Traffic state prediction based on spatiotemporal data has become a prominent focus in data-driven AI research. While significant progress has been made, most mainstream approaches assume uniform spatial and temporal correlations across conditions and use shared parameters for all scenarios. This simplification overlooks the complexity and heterogeneity inherent in human mobility patterns, often leading to suboptimal predictions. Recently, methods adopting the "decompose, then predict" (DTP) paradigm have gained traction. These methods break down data into smaller, manageable subcomponents, each predicted using dedicated parameters. Although effective in practice, DTP methods face unresolved theoretical questions: What type of decomposition truly makes subcomponents more manageable than the original data? To address this, we present an information theory-based analysis that derives sufficient conditions for a decomposition algorithm to reduce data-induced prediction errors. These conditions suggest that an effective algorithm should ensure decomposed components are as independent as possible, a principle we term the Component Independence Principle. Guided by this principle, we introduce the Theory-guided Graph Decomposition Learning (TGDL) framework, which decomposes graph-based multivariate time series data into approximately independent subgraph components that are easier to predict than the original data. Moreover, TGDL is a portable framework that can be integrated into any graph-based traffic prediction model to improve its predictive performance. Extensive experiments on four public datasets demonstrate the effectiveness of our approach. With a solid theoretical foundation, our TGDL enhances the performance of diverse traffic prediction models, yielding an average improvement of 19.37% across experiments.

Recent advances in machine learning (ML) have raised growing privacy concerns. One notable concern of ML models is the membership leakage risk, which is defined as the extent to which an adversary can accurately determine whether a record was part of a model's training set through membership inference attacks (MIAs). While extensively studied in non-time-series models, this risk remains largely unexplored for time-series prediction models. To address this gap, this work aims to provide a precise evaluation of membership leakage risks for time-series prediction models. We first conduct an empirical analysis to identify distinct temporal discrepancies within sequential outputs between member and non-member samples in time-series models. Beyond global temporal discrepancies, our findings show that member records exhibit notable local discrepancies, including smaller fluctuations in their output sequences when actual values remain constant and more rapid adjustments when actual values change, compared with non-member records. Building on these insights, we propose TSP-MIA, a framework that leverages a patch-based attack model to capture both global and local discrepancies for more effective MIAs. Extensive evaluations demonstrate the effectiveness and robustness of TSP-MIA, revealing substantial membership leakage risks for time-series prediction models. We further assess the validity of existing defense mechanisms against TSP-MIA.

Generalizing to unseen graph tasks without task-specific supervision remains challenging.Graph Neural Networks (GNNs) are limited by fixed label spaces, while Large Language Models (LLMs) lack structural inductive biases.Recent advances in Large Reasoning Models (LRMs) provide a zero-shot alternative via explicit, long chain-of-thought reasoning.Inspired by this, we propose a GNN-free approach that reformulates graph tasks-node classification, link prediction, and graph classification-as textual reasoning problems solved by LRMs.We introduce the first datasets with detailed reasoning traces for these tasks and develop GRAPH-R1, a reinforcement learning framework that leverages task-specific rethink templates to guide reasoning over linearized graphs.Experiments demonstrate that GRAPH-R1 outperforms state-of-the-art baselines in zero-shot settings, producing interpretable and effective predictions.Our work highlights the promise of explicit reasoning for graph learning and provides new resources for future research.Codes are available at https://github.com/lgybuaa/Graph-R1.

Recent years have witnessed the perfect encounter of deep learning and quantitative trading has achieved great success in stock investment. Numerous deep learning-based models have been developed for forecasting stock returns, leveraging the powerful representation capabilities of neural networks to identify patterns and factors influencing stock prices. These models can effectively capture general patterns in the market, such as stock price trends, volume-price relationships, and time variations. However, the impact of special irrationality factors -- such as market sentiment, speculative behavior, market manipulation, and psychological biases -- has not been fully considered in existing deep stock forecasting models due to their relative abstraction as well as lack of explicit labels and data description. To fill this gap, we propose UMI, a Universal multi-level Market Irrationality factor model to enhance stock return forecasting. The UMI model learns factors that can reflect irrational behaviors in market from both individual stock and overall market levels. For the stock-level, UMI construct an estimated rational price for each stock, which is cointegrated with the stock's actual price. The discrepancy between the actual and the rational prices serves as a factor to indicate stock-level irrational events. Additionally, we define market-level irrational behaviors as anomalous synchronous fluctuations of stocks within a market. Using two self-supervised representation learning tasks, i.e., sub-market comparative learning and market synchronism prediction, the UMI model incorporates market-level irrationalities into a market representation vector, which is then used as the market-level irrationality factor. We also developed a forecasting model that captures both temporal and relational dependencies among stocks, accommodating the UMI factors. Extensive experiments on U.S. and Chinese stock markets with competitive baselines demonstrate our model's effectiveness and the universality of our factors in improving various forecasting models. We provide our code at https://github.com/lIcIIl/UMI.

Generalizing to unseen graph tasks without task-specific supervision is challenging: conventional graph neural networks are typically tied to a fixed label space, while large language models (LLMs) struggle to capture graph structure. We introduce UniGTE, an instruction-tuned encoder–decoder framework that unifies structural and semantic reasoning. The encoder augments a pretrained autoregressive LLM with learnable alignment tokens and a structure-aware graph–text attention mechanism, enabling it to attend jointly to a tokenized graph and a natural-language task prompt while remaining permutation-invariant to node order. This yields compact, task-aware graph representations. Conditioned solely on these representations, a frozen LLM decoder predicts and reconstructs: it outputs the task answer and simultaneously paraphrases the input graph in natural language. The reconstruction objective regularizes the encoder to preserve structural cues. UniGTE is instruction-tuned on five datasets spanning node-, edge-, and graph-level tasks across diverse domains, yet requires no fine-tuning at inference. It achieves new state-of-the-art zero-shot results on node classification, link prediction, graph classification and graph regression under cross-task and cross-domain settings, demonstrating that tight integration of graph structure with LLM semantics enables robust, transferable graph reasoning.

With the increasing use of recommender systems in various application domains, many algorithms have been proposed for improving the accuracy of recommendations. Among various dimensions of recommender systems performance, long-tail (niche) recommendation performance remains an important challenge in large part because of the popularity bias of many existing recommendation techniques. In this study, we propose CORE, a cosine pattern–based technique, for effective long-tail recommendation. Comprehensive experimental results compare the proposed approach with a wide variety of classic, widely used recommendation algorithms and demonstrate its practical benefits in accuracy, flexibility, and scalability in addition to the superior long-tail recommendation performance. 1 History: Accepted by Ramaswamy Ramesh, Area Editor for Data Science & Machine Learning. Funding: This work was supported by the National Natural Science Foundation of China [Grants 72031001, 72072091, 72242101]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.0194 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2022.0194 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .

Publishing graph data is widely desired to enable a variety of structural analyses and downstream tasks. However, it also potentially poses severe privacy leakage, as attackers may leverage the released graph data to launch attacks and precisely infer private information such as the existence of hidden sensitive links in the graph. Prior studies on privacy-preserving graph data publishing relied on heuristic graph modification strategies and it is difficult to determine the graph with the optimal privacy–utility trade-off for publishing. In contrast, we propose the first privacy-preserving graph structure learning framework against sensitive link inference attacks, named PPGSL, which can automatically learn a graph with the optimal privacy–utility trade-off. The PPGSL operates by first simulating a powerful surrogate attacker conducting sensitive link attacks on a given graph. It then trains a parameterized graph to defend against the simulated adversarial attacks while maintaining the favorable utility of the original graph. To learn the parameters of both parts of the PPGSL, we introduce a secure iterative training protocol. It can enhance privacy preservation and ensure stable convergence during the training process, as supported by the theoretical proof. Additionally, we incorporate multiple acceleration techniques to improve the efficiency of the PPGSL in handling large-scale graphs. The experimental results confirm that the PPGSL achieves state-of-the-art privacy–utility trade-off performance and effectively thwarts various sensitive link inference attacks.

Fraud detection has always been one of the primary concerns in social and economic activities and is becoming a decisive force in the booming digital economy. Graph structures formed by rich user interactions naturally serve as important clues for identifying fraudsters. While numerous graph neural network-based methods have been proposed, the diverse interactive connections within graphs and the heterophilic connections deliberately established by fraudsters to normal users as camouflage pose new research challenges. In this light, we propose H2IDE (Homophily and Heterophily Identification with Disentangled Embeddings) for accurate fraud detection in multi-relation graphs. H2IDE features in an independence-constrained disentangled representation learning scheme to capture various latent behavioral patterns in graphs, along with a supervised identification task to specifically model the factor-wise heterophilic connections, both of which are proven crucial to fraud detection. We also design a relation-aware attention mechanism for hierarchical and adaptive neighborhood aggregation in H2IDE. Extensive comparative experiments with state-of-the-art baseline methods on two real-world multi-relation graphs and two large-scale homogeneous graphs demonstrate the superiority and scalability of our proposed method and highlight the key role of disentangled representation learning with homophily and heterophily identification.

Zero-shot graph machine learning, especially with graph neural networks (GNNs), has garnered significant interest due to the challenge of scarce labeled data. While methods like self-supervised learning and graph prompt learning have been extensively explored, they often rely on fine-tuning with task-specific labels, limiting their effectiveness in zero-shot scenarios. Inspired by the zero-shot capabilities of instruction-fine-tuned large language models (LLMs), we introduce a novel framework named Token Embedding-Aligned Graph Language Model (TEA-GLM) that leverages LLMs as cross-dataset and cross-task zero-shot learners for graph machine learning. Concretely, we pretrain a GNN, aligning its representations with token embeddings of an LLM. We then train a linear projector that transforms the GNN's representations into a fixed number of graph token embeddings without tuning the LLM. A unified instruction is designed for various graph tasks at different levels, such as node classification (node-level) and link prediction (edge-level). These design choices collectively enhance our method's effectiveness in zero-shot learning, setting it apart from existing methods. Experiments show that our graph token embeddings help the LLM predictor achieve state-of-the-art performance on unseen datasets and tasks compared to other methods using LLMs as predictors. Our code is available at https://github.com/W-rudder/TEA-GLM.

A gray market emerges when some distributors divert products to unauthorized distributors/retailers to make sneaky profits from the manufacturers’ differential channel incentives, such as quantity discounts. Traditionally, manufacturers rely heavily on internal audits to periodically investigate the flows of products and funds so as to deter the gray market; however, this is too costly given the large number of distributors and their huge volumes of orders. Owing to the advances in data analytics techniques, the ordering quantities of a distributor over time, which form multivariate time series, can help reveal suspicious product diversion behaviors and narrow the audit scope drastically. To that end, in this paper, we build on the recent advancement of representation learning for time series and adopt a sequence autoencoder to automatically characterize the overall demand patterns. To cope with the underlying entangled factors and interfering information in the multivariate time series of ordering quantities, we develop a disentangled learning scheme to construct more effective sequence representations. An interdistributor correlation regularization is also proposed to ensure more reliable representations. Finally, given the highly scarce anomaly labels for the detection task, an unsupervised deep generative model based on the learned representations of the distributors is developed to estimate the densities of distributions, which enables the anomaly scores generated through end-to-end learning. Extensive experiments on a real-world distribution channel data set and a larger simulated data set empirically validate our model’s superior and robust performances compared with several state-of-the-art baselines. Additionally, our illustrative economic analysis demonstrates that the manufacturers can launch more targeted and cost-effective audits toward the suspected distributors recommended by our model so as to deter the gray market. History: Accepted by Ram Ramesh, Area Editor for Data Science & Machine Learning. Funding: This work was supported by the National Natural Science Foundation of China [Grants 72031001, 72301017, 72371011, and 72242101]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.0155 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2022.0155 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .

Network embedding represents network nodes by a low-dimensional informative vector. While it is generally effective for various downstream tasks, it may leak some private information of networks, such as hidden private links. In this work, we address a novel problem ofprivacy-preserving network embedding against private link inference attacks. Basically, we propose to perturb the original network by adding or removing links, and expect the embedding generated on the perturbed network can leak little information about private links but hold high utility for various downstream tasks. Towards this goal, we first propose general measurements to quantify privacy gain and utility loss incurred by candidate network perturbations; we then design aPrivacy-PreservingNetworkEmbedding (i.e., PPNE) framework to identify the optimal perturbation solution with the best privacy-utility trade-off in an iterative way. Furthermore, we propose many techniques to accelerate PPNE and ensure its scalability. For instance, as the skip-gram embedding methods including DeepWalk and LINE can be seen as matrix factorization with closed-form embedding results, we devise efficient privacy gain and utility loss approximation methods to avoid the repetitive time-consuming embedding training for every candidate network perturbation in each iteration. Experiments on real-life network datasets (with up to millions of nodes) verify that PPNE outperforms baselines by sacrificing less utility and obtaining higher privacy protection.

Minimizing privacy leakage while ensuring data utility is a critical problem in a privacy-preserving data publishing task, from which data holders can boost platform engagements or enlarge data values. Most prior research concerned only with either privacy-insensitive or exact private data and resorts to a single obscuring method to achieve a privacy-utility tradeoff, which is inadequate for real-life hybrid data especially when facing machine learning-based inference attacks. This work takes a pilot study on privacy-preserving data publishing when both widely adopted generalization and obfuscation operations are employed for privacy-heterogeneous data protection. Specifically, we first propose novel measures for privacy and utility values quantification and formulate the hybrid privacy-preserving data obscuring problem to account for the joint effect of generalization and obfuscation. We then design a novel protection mechanism called HyObscure, which decomposes the original problem into three sub-problems to cross-iteratively optimize the hybrid operations for maximum privacy protection under a certain data utility guarantee. The convergence of the iterative process and the privacy leakage bound of HyObscure are also provided in theory. Extensive experiments demonstrate that HyObscure significantly outperforms a variety of state-of-the-art baseline methods when facing various inference attacks in different scenarios.

This paper presents a novel map matching framework that adopts deep learning techniques to map a sequence of cell tower locations to a trajectory on a road network. Map matching is an essential pre-processing step for many applications, such as traffic optimization and human mobility analysis. However, most recent approaches are based on hidden Markov models (HMMs) or neural networks that are hard to consider high-order location information or heuristics observed from real driving scenarios. In this paper, we develop a deep reinforcement learning based map matching framework for cellular data, named as DMM, which adopts a recurrent neural network (RNN) coupled with a reinforcement learning scheme to identify the most-likely trajectory of roads given a sequence of cell towers. To transform DMM into a practical system, several challenges are addressed by developing a set of techniques, including spatial-aware representation of input cell tower sequences, an encoder-decoder based RNN network for map matching model with variable-length input and output, and a global heuristics-driven reinforcement learning based scheme for optimizing the parameters of the encoder-decoder map matching model. Extensive experiments on a large-scale anonymized cellular dataset reveal that DMM provides high map matching accuracy and fast inference time.

Recent years have witnessed the growth of Graph-based Collaborative Filtering (GCF) for high-performance recommendations, but the widely adopted user-item bipartite graphs are subject to deeper layers' over-smoothing effect and sparse user-item interactions when learning item representations. In this work, we introduceitem graph, which regards items as nodes and connecting those that have ever co-occurred in transactions with edges, to preserve higher-order item relations while avoiding the drawbacks of bipartite graphs for item-based recommendation. To cope with the entangled semantics in the edges of an item graph, we first design a denoising scheme via a graph structure learning module with discrete sampling to drop noisy edges with respect to certain latent aspects, where multiple subgraphs can be yielded. We then implement graphical disentangled learning by imposing several structural regularizers that allow for macro conformity and micro divergence among the subgraphs. Finally, we propose a multi-graph fusion module to aggregate users' preferences in different subgraphs with a user-graph attention mechanism. Extensive experiments on 5 real-world datasets demonstrate the superiority of our method over 16 competitive baseline methods including the recently proposed GCF ones. Particularly, our method shows evident advantages in recommendation under data sparsity conditions.

Multilabel learning involving hundreds of thousands or even millions of labels is referred to as extreme multilabel learning (XML), in which the labels often follow a power-law distribution with the majority occurring in very few data points as tail labels. Recent years have witnessed the intensive use of deep-learning methods for high-performance XML, but they are typically optimized for the head labels with abundant training instances and less consider the performance on tail labels, which, however, like the needles in haystacks, are often the focus of attention in real-life applications. In light of this, we present BoostXML, a deep learning-based XML method for extreme multilabel text classification, enhanced greatly by gradient boosting. In BoostXML, we pay more attention to tail labels in each Boosting Step by optimizing the residual mostly from unfitted training instances with tail labels. A Corrective Step is further proposed to avoid the mismatching between the text encoder and weak learners during optimization, which reduces the risk of falling into local optima and improves model performance. A Pretraining Step is also introduced in the initial stage of BoostXML to avoid exorbitant bias to tail labels. Extensive experiments on five benchmark datasets with state-of-the-art baselines demonstrate the advantage of BoostXML in tail-label prediction.

Robust prediction of citywide traffic flows at different time periods plays a crucial role in intelligent transportation systems. While previous work has made great efforts to model spatio-temporal correlations, existing methods still suffer from two key limitations: i) Most models collectively predict all regions' flows without accounting for spatial heterogeneity, i.e., different regions may have skewed traffic flow distributions. ii) These models fail to capture the temporal heterogeneity induced by time-varying traffic patterns, as they typically model temporal correlations with a shared parameterized space for all time periods. To tackle these challenges, we propose a novel Spatio-Temporal Self-Supervised Learning (ST-SSL) traffic prediction framework which enhances the traffic pattern representations to be reflective of both spatial and temporal heterogeneity, with auxiliary self-supervised learning paradigms. Specifically, our ST-SSL is built over an integrated module with temporal and spatial convolutions for encoding the information across space and time. To achieve the adaptive spatio-temporal self-supervised learning, our ST-SSL first performs the adaptive augmentation over the traffic flow graph data at both attribute- and structure-levels. On top of the augmented traffic graph, two SSL auxiliary tasks are constructed to supplement the main traffic prediction task with spatial and temporal heterogeneity-aware augmentation. Experiments on four benchmark datasets demonstrate that ST-SSL consistently outperforms various state-of-the-art baselines. Since spatio-temporal heterogeneity widely exists in practical datasets, the proposed framework may also cast light on other spatial-temporal applications. Model implementation is available at https://github.com/Echo-Ji/ST-SSL.

Given its broad applications, time series analysis has gained substantial research attention but remains a very challenging task. Recent years have witnessed the great success of deep learning methods, eg., CNN and RNN, in time series classification and forecasting, but heterogeneity as the very nature of time series has not yet been addressed adequately and remains the performance "treadstone." In this light, we argue that the intra-sequence non-stationarity and inter-sequence asynchronism are two types of heterogeneities widely existed in multiple times series, and propose a hybrid attention network called WHEN as deep learning solution. WHEN features in two attention mechanisms in two different modules. In the WaveAtt module, we propose a novel data-dependent wavelet function and integrate it into the BiLSTM network as the wavelet attention, for the purpose of analyzing dynamic frequency components in nonstationary time series. In the DTWAtt module, we transform the dynamic time warping (DTW) technique into the form as the DTW attention, where all input sequences are synchronized with a universal parameter sequence to overcome the time distortion problem in multiple time series. WHEN with the hybrid attentions is then formulated as task-dependent neural network for either classification or forecasting tasks. Extensive experiments on 30 UEA datasets and 3 real-world datasets with rich competitive baselines demonstrate the excellent performance of our model. The ability of WHEN in dealing with time series heterogeneities is also elaborately explored via specially designed analysis.

Next Basket Recommendation (NBR) that recommends a basket of items to users has become a promising promotion artifice for online businesses. The key challenge of NBR is rooted in the complicated relations of items that are dependent on one another in a same basket with users' diverse purchasing intentions, which goes far beyond the pairwise item relations in traditional recommendation tasks, and yet has not been well addressed by existing NBR methods that mostly model the inter-basket item relations only. To that end, in this paper, we construct a hypergraph from basket-wise purchasing records and probe the inter-basket and intra-basket item relations behind the hyperedges. In particular, we combine the strength of HyperGraph Neural Network with disentangled representation learning to derive the intent-aware representations of hyperedges for characterizing the nuances of user purchasing patterns. Moreover, considering the information loss in traditional item-wise optimization, we propose a novel basket-wise optimization scheme via an adversarial network to generate high-quality negative baskets. Extensive experiments conducted on four different data sets demonstrate the superior performances over the state-of-the-art NBR methods. Notably, our method is shown to strike a good balance in recommending both repeated and explorative items as a basket.

Detecting frauds from a massive amount of user behavioral data is often regarded as finding a needle in a haystack. While tremendous efforts have been devoted to fraud detection from behavioral sequences, existing studies rarely consider behavioral targets and companions and their interactions simultaneously in a sequence model. In this paper, we suggest extracting source and target neighbor sequences from the temporal bipartite network of user behaviors, and disclose the interesting correlation mode and repetition mode hidden inside the two types of sequences as important clues for fraudsters distinguishment. We then propose a novel Hawkes-enhanced sequence model (HESM) by integrating the Hawkes process into LSTM for historical influence learning. A historical attention mechanism is also proposed to enhance the strength of the long-term historical influence in response to the repetition mode. Moreover, in order to collectively model both types of neighbor sequences for capturing the correlation mode, we propose a correlation gate to control the information flow in sequences. We conduct extensive experiments on real-world datasets and demonstrate that HESM outperforms competitive baseline methods consistently in telecom fraud detection. Particularly, the abilities of HESM in historical influence leaning and sequence correlation learning have been explored visually and intensively.

Given the effectiveness and ease of use, Item-based Collaborative Filtering (ICF) methods have been broadly used in industry and are widely investigated in recent years. The key of ICF lies in the similarity measurement between items, which however is a coarse-grained numerical value that can hardly capture users' fine-grained preferences toward different attributed aspects of items. In this paper, we propose a model called REDA to address this challenge, based on which a new paradigm called Relation-based Collaborative Filtering is designed for high-performance recommendation. REDA is essentially a deep neural network model that employs an item relation embedding scheme for inter-item relations representation. It features in multi-decomposed item embedding with dual-attention refinement and employs a novel relation-wise optimization scheme for end-to-end learning. A relational user embedding is then proposed by aggregating item relation embeddings between all purchased items of a user, which not only profiles users' fine-grained preferences but also alleviates the data sparsity problem. Extensive experiments are conducted on five real-world datasets and the proposed REDA is shown to outperform ten state-of-the-art methods. In particular, REDA shows great robustness against data and relation sparsity, the ability to learn explainable item aspects, and the potential for large-scale recommendation.

Existing studies have shown that user-item interaction data and social relation data can be jointly used for enhancing the performance of social recommendation. However, limited research has a focus on investigating how to deeply exploit different views of social interaction structures and rating behavior differences for further improving social recommendation. To this end, in this paper, we propose to integrate information from both semantic and structural views for social recommendation. Specifically, we first design a collective intelligence-based strategy to reveal high-quality implicit relations for both users and items. Then, by reformulating all available nodes and relations as a heterogeneous graph, we define multiple semantic metapaths to capture diverse preferences for comprehensive user and item representations. While various metapaths enlarge the representation capacity of users and items, they also introduce noise and irrelevant information. We recall that, for the user-item interaction graph, different structure sizes (e.g., local and global structures) provide diverse and complementary information for recommendation. Motivated by this, we propose a semantic and structural view fusion framework for social recommendation (S4Rec), which consists of a deep graph model and a wide attentive SVD (Singular Value Decomposition) model for rating prediction by taking the local and global structure as input and aggregating messages along the predefined metapaths. Finally, the two predicted results are adaptively fused to achieve the final both accurate and stable prediction. In addition, we treat the user's rating behavior difference as the relative position difference problem in the embedding space, and model it with TransH to improve the generalization ability of the main rating model. Extensive experiments on three open datasets demonstrate the superiority of our framework compared with state-of-the-art methods. Particularly, our model outperforms other baselines under different sparsity conditions, further validating the effectiveness on cold-start users. We release the source code athttps://github.com/lcwy220/Social-Recommendation.

Recent years have witnessed the unprecedented growth of online social media, resulting in short texts being the prevalent format of information on the Internet. Given the sparsity of data, however, short-text topic modeling remains a critical yet much-watched challenge in both academia and industry. Research has been devoted to building different types of probabilistic topic models for short texts, among which self-aggregation methods emerged recently to provide informative cross-text word co-occurrences. However, models along this line are still in their infancy and typically yield overfit results and exhibit high computational costs. In this paper, we propose a novel model called Pseudo-document-based Topic Model (PTM), which introduces the concept of pseudo-document to implicitly aggregate short texts against data sparsity. By modeling the topic distributions of latent pseudo-documents rather than short texts, PTM yields excellent performance in accuracy and efficiency. A word embedding-enhanced PTM (WE-PTM) is also proposed to leverage pre-trained word embeddings, which is essential to further alleviating data sparsity. Extensive experiments with self-aggregation or word embedding-based baselines on four real-world datasets including two online media short texts, demonstrate the high-quality topics learned by our models. Robustness to limited training samples and the explainable semantics of topics are also investigated.

Consensus clustering is gaining increasing attention for its high quality and robustness. In particular, k -means-based Consensus Clustering (KCC) converts the usual computationally expensive problem to a classic k -means clustering with generalized utility functions, bringing potentials for large-scale data clustering on different types of data. Despite KCC’s applicability and generalizability, implementing this method such as representing the binary dataset in the k -means heuristic is challenging and has seldom been discussed in prior work. To fill this gap, we present a MATLAB package, KCC, that completely implements the KCC framework and utilizes a sparse representation technique to achieve a low space complexity. Compared to alternative consensus clustering packages, the KCC package is of high flexibility, efficiency, and effectiveness. Extensive numerical experiments are also included to show its usability on real-world datasets.

With the unprecedented development of industrialization and urbanization, many hazardous chemicals have become an indispensable part of our daily life. They are produced, transported, and consumed in modern cities every day, which breeds many unknown hazardous chemicals-related locations (HCLs) that are out of the supervision of management departments and accompanying huge threats to urban safety. How to recognize these unknown HCLs and identify their risk levels is an essential task for urban hazardous chemicals management. To accomplish this task, in this work, we propose a system named as CityShield to discover hidden HCLs and classify their risk levels based on trajectories of hazardous chemicals transportation vehicles. The CityShield system consists of three components. The first component is Data Pre-processing, which filters noises in raw trajectories and probes stable transportation vehicles' stay points from massive uncertain GPS points. The second is HCL Recognition, which adopts the proposed HCL-Rec algorithm to cluster stay points into polygonal HCLs, and avoids the improper location merging problem caused by the skewed spatial distribution of HCLs. The third component is HCL Classification, which introduces the HCL relation graph as auxiliary information to overcome the label scarcity problem of HCLs. It adopts a self-supervised method consisting of four pre-training tasks to learn high-quality representations for HCLs from the graph, which are finally used to classify the categories and risk levels of HCLs.

Aspect Sentiment Triplet Extraction (ASTE) is a relatively new and very challenging task that attempts to provide an integral solution for aspect-based sentiment analysis. Aspect sentiment triplets in a sentence usually have overlaps when, e.g., one aspect is associated with multiple opinions and vice versa. Recently, end-to-end ASTE methods are becoming more and more popular for they can avoid the error propagation problem of pipeline-based methods. However, existing tagging-based end-to-end methods face difficulty to obtain a satisfactory recall, and generative methods fail to take a full account of the underlying interactions between aspects, opinions and their corresponding sentiments. In this paper, we formalize the ASTE task as a Seq2Seq learning problem with span copy mechanism for extracting multiple and possibly overlapped triplets. A novel dual decoder is devised purposefully for the ASTE task, where a multi-head attention based span copy mechanism is proposed to copy multi-token aspects and opinions. The dual decoder benefits from the rich output of encoder that can fuse multi-type information including word semantic, POS tag and BIO tag. Experiments on various benchmark datasets demonstrate that our approach achieves new state-of-the-art results. We also conduct analytical experiments to verify the effectiveness of various model components particularly for overlapped triplets extraction. We find that our model can be further improved through data augmentation and post-training.

The cross-lingual sentiment analysis (CLSA) aims to leverage label-rich resources in the source language to improve the models of a resource-scarce domain in the target language, where monolingual approaches based on machine learning usually suffer from the unavailability of sentiment knowledge. Recently, the transfer learning paradigm that can transfer sentiment knowledge from resource-rich languages, for example, English, to resource-poor languages, for example, Chinese, has gained particular interest. Along this line, in this article, we propose semisupervised learning with SCL and space transfer (ssSCL-ST), a semisupervised transfer learning approach that makes use of structural correspondence learning as well as space transfer for cross-lingual sentiment analysis. The key idea behind ssSCL-ST, at a high level, is to explore the intrinsic sentiment knowledge in the target-lingual domain and to reduce the loss of valuable knowledge due to the knowledge transfer via semisupervised learning. ssSCL-ST also features in pivot set extension and space transfer, which helps to enhance the efficiency of knowledge transfer and improve the classification accuracy in the target language domain. Extensive experimental results demonstrate the superiority of ssSCL-ST to the state-of-the-art approaches without using any parallel corpora.

The recent pandemic has painfully reminded us of the need for IS design to be informed by possible futures and conscious of undesirable futures – Within months, many of the nice-to-have IS functionalities have become must-haves; Technology solutions in response to the pandemic have raised privacy and equality concerns. Although design science research fundamentally focuses on shaping artefacts and events to create a more desirable future, there has been limited guidance on how futures should be accounted for. This article addresses the gap by integrating insights from future-oriented IS research and futures research to develop guidelines for engaging with futures throughout the design science research process. The future-ready design research (FRDR) process prompts researchers to be more aware of futures, to foster the innovative foresight for actively pursuing the preferred future, and to espouse the responsible foresight for consciously avoiding undesirable futures. The guidelines are illustrated with a design science research project on outbreak analytics and the instantiated system’s subsequent adaptation and utilisation in COVID-19.

The fuzzy cognitive map (FCM) is a powerful model for system state prediction and interpretable knowledge representation. Recent years have witnessed the tremendous efforts devoted to enhancing the basic FCM, such as introducing temporal factors, uncertainty or fuzzy rules to improve interpretation, and introducing fuzzy neural networks or wavelets to improve time series prediction. But how to achieve high-precision yet interpretable prediction in cross-domain real-life applications remains a great challenge. In this article, we propose a novel FCM extension called deep FCM (DFCM) for multivariate time series forecasting, in order to take both the advantage of FCM in interpretation and the advantage of deep neural networks in prediction. Specifically, to improve the predictive power, DFCM leverages a fully connected neural network to model connections (relationships) among concepts in a system, and a recurrent neural network to model unknown exogenous factors that have influences on system dynamics. Moreover, to foster model interpretability encumbered by the embedded deep structures, a partial derivative-based approach is proposed to measure the connection strengths between concepts in DFCM. An alternate function gradient descent algorithm is then proposed for parameter inference. The effectiveness of DFCM is validated over four publicly available datasets with the presence of seven baselines. DFCM indeed provides an important clue to building interpretable predictors for real-life applications.

Cross-lingual sentiment classification (CLSC) aims to leverage rich-labeled resources in the source language to improve prediction models of a resource-scarce domain in the target language. Existing feature representation learning-based approaches try to minimize the difference of latent features between different domains by exact alignment, which is achieved by either one-to-one topic alignment or matrix projection. Exact alignment, however, restricts the representation flexibility and further degrades the model performances on CLSC tasks if the distribution difference between two language domains is large. On the other hand, most previous studies proposed document-level models or ignored sentiment polarities of topics that might lead to insufficient learning of latent features. To solve the abovementioned problems, we propose a coarse alignment mechanism to enhance the model's representation by a group-to-group topic alignment into an aspect-level fine-grained model. First, we propose an unsupervised aspect, opinion, and sentiment unification model (AOS), which trimodels aspects, opinions, and sentiments of reviews from different domains and helps capture more accurate latent feature representation by a coarse alignment mechanism. To further boost AOS, we propose ps-AOS, a partial supervised AOS model, in which labeled source language data help minimize the difference of feature representations between two language domains with the help of logistics regression. Finally, an expectation-maximization framework with Gibbs sampling is then proposed to optimize our model. Extensive experiments on various multilingual product review data sets show that ps-AOS significantly outperforms various kinds of state-of-the-art baselines.

Recent years have witnessed the emergence of worldwide megalopolises and the accompanying public safety events, making urban safety a top priority in modern urban management. Among various threats, dangerous goods such as gas and hazardous chemicals transported through cities have bred repeated tragedies and become the deadly “bomb” we sleep with every day. While tremendous research efforts have been devoted to dealing with dangerous goods transportation (DGT) issues, further study is still in great need to quantify this problem and explore its intrinsic dynamics from a big data perspective. In this article, we present a novel system called DGeye , to feature a fusion between DGT trajectory data and residential population data for dangers perception and prediction. Specifically, DGeye first develops a probabilistic graphical model-based approach to mine spatio-temporally adjacent risk patterns from population-aware risk trajectories. Then, DGeye builds the novel causality network among risk patterns for risk pain-point identification, risk source attribution, and online risky state prediction. Experiments on both Beijing and Tianjin cities demonstrate the effectiveness of DGeye in real-life DGT risk management. As a case in point, our report powered by DGeye successfully drove the government to lay down gas pipelines for the famous Guijie food street in Beijing.

The main challenge of cross-domain text classification is to train a classifier in a source domain while applying it to a different target domain. Many transfer learning-based algorithms, for example, dual transfer learning, triplex transfer learning, etc., have been proposed for cross-domain classification, by detecting a shared low-dimensional feature representation for both source and target domains. These methods, however, often assume that the word clusters matrix or the clusters association matrix as knowledge transferring bridges are exactly the same across different domains, which is actually unrealistic in real-world applications and, therefore, could degrade classification performance. In light of this, in this paper, we propose a softly associative transfer learning algorithm for cross-domain text classification. Specifically, we integrate two non-negative matrix tri-factorizations into a joint optimization framework, with approximate constraints on both word clusters matrices and clusters association matrices so as to allow proper diversity in knowledge transfer, and with another approximate constraint on class labels in source domains in order to handle noisy labels. An iterative algorithm is then proposed to solve the above problem, with its convergence verified theoretically and empirically. Extensive experimental results on various text datasets demonstrate the effectiveness of our algorithm, even with the presence of abundant state-of-the-art competitors.

Spammers, who manipulate online reviews to promote or suppress products, are flooding in online commerce. To combat this trend, there has been a great deal of research focused on detecting review spammers, most of which design diversified features and thus develop various classifiers. The widespread growth of crowdsourcing platforms has created large-scale deceptive review writers who behave more like normal users, that the way they can more easily evade detection by the classifiers that are purely based on fixed characteristics. In this paper, we propose a hybrid semisupervised learning model titled hybrid PU-learning-based spammer detection (hPSD) for spammer detection to leverage both the users' characteristics and the user-product relations. Specifically, the hPSD model can iteratively detect multitype spammers by injecting different positive samples, and allows the construction of classifiers in a semisupervised hybrid learning framework. Comprehensive experiments on movie dataset with shilling injection confirm the superior performance of hPSD over existing baseline methods. The hPSD is then utilized to detect the hidden spammers from real-life Amazon data. A set of spammers and their underlying employers (e.g., book publishers) are successfully discovered and validated. These demonstrate that hPSD meets the real-world application scenarios and can thus effectively detect the potentially deceptive review writers.

The correlation network, typically formed by computing pairwise correlations between variables, has recently become a competitive paradigm to discover insights in various application domains, such as climate prediction, financial marketing, and bioinformatics. In this study, we adopt this paradigm to detect cheating behavior hidden in business distribution channels, where falsified big deals are often made by collusive partners to obtain lower product prices—a behavior deemed to be extremely harmful to the sale ecosystem. To this end, we assume that abnormal deals are likely to occur between two partners if their purchase-volume sequences have a strong negative correlation. This seemingly intuitive rule, however, imposes several research challenges. First, existing correlation measures are usually symmetric and thus cannot distinguish the different roles of partners in cheating. Second, the tick-to-tick correspondence between two sequences might be violated due to the possible delay of purchase behavior, which should also be captured by correlation measures. Finally, the fact that any pair of sequences could be correlated may result in a number of false-positive cheating pairs, which need to be corrected in a systematic manner. To address these issues, we propose a correlation network analysis framework for cheating detection. In the framework, we adopt an asymmetric correlation measure to distinguish the two roles, namely, cheating seller and cheating buyer , in a cheating alliance. Dynamic Time Warping is employed to address the time offset between two sequences in computing the correlation. We further propose two graph-cut methods to convert the correlation network into a bipartite graph to rank cheating partners, which simultaneously helps to remove false-positive correlation pairs. Based on a 4-year real-world channel dataset from a worldwide IT company, we demonstrate the effectiveness of the proposed method in comparison to competitive baseline methods.

It is always deemed crucial to identify the key factors that could have significant impact on the stock market trend. Recently, an interesting phenomenon has emerged that some of President Trump’s posts in Twitter can surge into a dominant role on the stock market for a certain time period, although studies along this line are still in their infancy. Therefore, in this article, we study whether and how this new-rising information can help boost the performance of stock market prediction. Specifically, we have found that the echoing reinforced effect of financial news with Trump’s market-related tweets can influence the market movement—that is, some of Trump’s tweets directly impact the stock market in a short time, and the impact can be further intensified when it echoes with other financial news reports. Along this line, we propose a deep information echoing model to predict the hourly stock market trend, such as the rise and fall of the Dow Jones Industrial Average. In particular, to model the discovered echoing reinforced impact, we design a novel information echoing module with a gating mechanism in a sequential deep learning framework to capture the fused knowledge from both Trump’s tweets and financial news. Extensive experiments have been conducted on the real-world U.S. stock market data to validate the effectiveness of our model and its interpretability in understanding the usability of Trump’s posts. Our proposed deep echoing model outperforms other baselines by achieving the best accuracy of 60.42% and obtains remarkable accumulated profits in a trading simulation, which confirms our assumption that Trump’s tweets contain indicative information for short-term market trends. Furthermore, we find that Trump’s tweets about trade and political events are more likely to be associated with short-term market movement, and it seems interesting that the impact would not degrade as time passes.

Lifelong topic model (LTM), an emerging paradigm for never-ending topic learning, aims to yield higher-quality topics as time passes through knowledge accumulated from the past yet learned for the future. In this paper, we propose a novel lifelong topic model based on non-negative matrix factorization (NMF), called Affinity Regularized NMF for LTM (NMF-LTM), which to our best knowledge is distinctive from the popular LDA-based LTMs. NMF-LTM achieves lifelong learning by introducing word-word graph Laplacian as semantic affinity regularization. Other priors such as sparsity, diversity, and between-class affinity are incorporated as well for better performance, and a theoretical guarantee is provided for the algorithmic convergence to a local minimum. Extensive experiments on various public corpora demonstrate the effectiveness of NMF-LTM, particularly its human-like behaviors in two carefully designed learning tasks and the ability in topic modeling of big data. A further exploration of semantic relatedness in knowledge graphs and a case study on a large-scale real-world corpus exhibit the strength of NMF-LTM in discovering high-quality topics in an efficient and robust way.

Fraud detection from massive user behaviors is often regarded as trying to find a needle in a haystack. In this paper, we suggest abnormal behavioral patterns can be better revealed if both sequential and interaction behaviors of users can be modeled simultaneously, which however has rarely been addressed in prior work. Along this line, we propose a COllective Sequence and INteraction (COSIN) model, in which the behavioral sequences and interactions between source and target users in a dynamic interaction network are modeled uniformly in a probabilistic graphical model. More specifically, the sequential schema is modeled with a hierarchical Hidden Markov Model, and meanwhile it is shifted to the interaction schema to generate the interaction counts through Poisson factorization. A hybrid Gibbs-Variational algorithm is then proposed for efficient parameter estimation of the COSIN model. We conduct extensive experiments on both synthetic and real-world telecom datasets in different scales, and the results show that the proposed model outperforms some competitive baseline methods and is scalable. A case is further presented to show the precious explainability of the model.

Recent years have witnessed the world-wide emergence of mega-metropolises with incredibly huge populations. Understanding residents mobility patterns, or urban dynamics, thus becomes crucial for building modern smart cities. In this paper, we propose a Neighbor-Regularized and context-aware Non-negative Tensor Factorization model (NR-cNTF) to discover interpretable urban dynamics from urban heterogeneous data. Different from many existing studies concerned with prediction tasks via tensor completion, NR-cNTF focuses on gaining urban managerial insights from spatial, temporal, and spatio-temporal patterns. This is enabled by high-quality Tucker factorizations regularized by both POI-based urban contexts and geographically neighboring relations. NR-cNTF is also capable of unveiling long-term evolutions of urban dynamics via a pipeline initialization approach. We apply NR-cNTF to a real-life data set containing rich taxi GPS trajectories and POI records of Beijing. The results indicate: 1) NR-cNTF accurately captures four kinds of city rhythms and seventeen spatial communities; 2) the rapid development of Beijing, epitomized by the CBD area, indeed intensifies the job-housing imbalance; 3) the southern areas with recent government investments have shown more healthy development tendency. Finally, NR-cNTF is compared with some baselines on traffic prediction, which further justifies the importance of urban contexts awareness and neighboring regulations.

Employer Brand Evaluation (EBE) is to understand an employer’s unique characteristics to identify competitive edges. Traditional approaches rely heavily on employers’ financial information, including financial reports and filings submitted to the Securities and Exchange Commission (SEC), which may not be readily available for private companies. Fortunately, online recruitment services provide a variety of employers’ information from their employees’ online ratings and comments, which enables EBE from an employee’s perspective. To this end, in this article, we propose a method named Company Profiling–based Collaborative Topic Regression (CPCTR) to collaboratively model both textual (i.e., reviews) and numerical information (i.e., salaries and ratings) for learning latent structural patterns of employer brands. With identified patterns, we can effectively conduct both qualitative opinion analysis and quantitative salary benchmarking. Moreover, a Gaussian processes--based extension, GPCTR, is proposed to capture the complex correlation among heterogeneous information. Extensive experiments are conducted on three real-world datasets to validate the effectiveness and generalizability of our methods in real-life applications. The results clearly show that our methods outperform state-of-the-art baselines and enable a comprehensive understanding of EBE.

Recent years have witnessed the successful marriage of finance innovations and AI techniques in various finance applications including quantitative trading (QT). Despite great research efforts devoted to leveraging deep learning (DL) methods for building better QT strategies, existing studies still face serious challenges especially from the side of finance, such as the balance of risk and return, the resistance to extreme loss, and the interpretability of strategies, which limit the application of DL-based strategies in real-life financial markets. In this work, we propose AlphaStock, a novel reinforcement learning (RL) based investment strategy enhanced by interpretable deep attention networks, to address the above challenges. Our main contributions are summarized as follows: i) We integrate deep attention networks with a Sharpe ratio-oriented reinforcement learning framework to achieve a risk-return balanced investment strategy; ii) We suggest modeling interrelationships among assets to avoid selection bias and develop a cross-asset attention mechanism; iii) To our best knowledge, this work is among the first to offer an interpretable investment strategy using deep reinforcement learning models. The experiments on long-periodic U.S. and Chinese markets demonstrate the effectiveness and robustness of AlphaStock over diverse market states. It turns out that AlphaStock tends to select the stocks as winners with high long-term growth, low volatility, high intrinsic value, and being undervalued recently.

Driven by the wave of urbanization in recent decades, the research topic about migrant behavior analysis draws great attention from both academia and the government. Nevertheless, subject to the cost of data collection and the lack of modeling methods, most of existing studies use only questionnaire surveys with sparse samples and non-individual level statistical data to achieve coarse-grained studies of migrant behaviors. In this paper, a partially supervised cross-domain deep learning model named CD-CNN is proposed for migrant/native recognition using mobile phone signaling data as behavioral features and questionnaire survey data as incomplete labels. Specifically, CD-CNN features in decomposing the mobile data into location domain and communication domain, and adopts a joint learning framework that combines two convolutional neural networks with a feature balancing scheme. Moreover, CD-CNN employs a three-step algorithm for training, in which the co-training step is of great value to partially supervised cross-domain learning. Comparative experiments on the city Wuxi demonstrate the high predictive power of CD-CNN. Two interesting applications further highlight the ability of CD-CNN for in-depth migrant behavioral analysis.

Fraud detection is usually regarded as finding a needle in haystack, which is a challenging task because fraudulences are buried in massive normal behaviors. Indeed, a fraudulent incident usually takes place in consecutive time steps to gain illegal benefits, which provides unique clues to probing frauds by considering a complete behavioral sequence, rather than detecting frauds from a snapshot of behaviors. Also, fraudulent behaviors may entail different parties, such that the interaction pattern between sources and targets can help distinguish frauds from normal behaviors. Therefore, in this paper, we model the attributed behavioral sequences generated from consecutive behaviors, in order to capture the sequential patterns, while those deviate from the pattern can be regarded as fraudulence. Considering the characteristics of behavioral sequence, we propose a novel model, HAInt-LSTM, by augmenting traditional LSTM with a modified forget gate where interval time between consecutive time steps are considered. Meanwhile, we employ a self-historical attention mechanism to allow for long-time dependencies, which can help identify repeated or cyclical appearances. In addition, we encode the source information as an interaction module to enhance the learning of behavioral sequences. To validate the effectiveness of the learned sequential behavior representations, we experiment on real-world telecommunication dataset under both supervised and unsupervised scenarios. Experimental results show that the learned representations can better identify fraudulent behaviors, and also show a clear cut with normal sequences in the lower dimensional embedding space through visualization. Last but not least, we visualize the weights of attention mechanism to provide rational interpretation of human behavioral periodicity.

Given the rich real-life applications of network mining as well as the surge of representation learning in recent years, network embedding has become the focal point of increasing research interests in both academic and industrial domains. Nevertheless, the complete temporal formation process of networks characterized by sequential interactive events between nodes has yet seldom been modeled in the existing studies, which calls for further research on the so-called temporal network embedding problem. In light of this, in this paper, we introduce the concept of neighborhood formation sequence to describe the evolution of a node, where temporal excitation effects exist between neighbors in the sequence, and thus we propose a Hawkes process based Temporal Network Embedding (HTNE) method. HTNE well integrates the Hawkes process into network embedding so as to capture the influence of historical neighbors on the current neighbors. In particular, the interactions of low-dimensional vectors are fed into the Hawkes process as base rate and temporal influence, respectively. In addition, attention mechanism is also integrated into HTNE to better determine the influence of historical neighbors on current neighbors of a node. Experiments on three large-scale real-life networks demonstrate that the embeddings learned from the proposed HTNE model achieve better performance than state-of-the-art methods in various tasks including node classification, link prediction, and embedding visualization. In particular, temporal recommendation based on arrival rate inferred from node embeddings shows excellent predictive power of the proposed model.

Aspect-based opinion mining is finding elaborate opinions towards a subject such as a product or an event. With explosive growth of opinionated texts on the Web, mining aspect-level opinions has become a promising means for online public opinion analysis. In particular, the boom of various types of online media provides diverse yet complementary information, bringing unprecedented opportunities for cross media aspect-opinion mining. Along this line, we propose CAMEL, a novel topic model for complementary aspect-based opinion mining across asymmetric collections. CAMEL gains information complementarity by modeling both common and specific aspects across collections, while keeping all the corresponding opinions for contrastive study. An auto-labeling scheme called AME is also proposed to help discriminate between aspect and opinion words without elaborative human labeling, which is further enhanced by adding word embedding-based similarity as a new feature. Moreover, CAMEL-DP, a nonparametric alternative to CAMEL is also proposed based on coupled Dirichlet Processes. Extensive experiments on real-world multi-collection reviews data demonstrate the superiority of our methods to competitive baselines. This is particularly true when the information shared by different collections becomes seriously fragmented. Finally, a case study on the public event “2014 Shanghai Stampede” demonstrates the practical value of CAMEL for real-world applications.

Company profiling is an analytical process to build an in-depth understanding of company's fundamental characteristics. It serves as an effective way to gain vital information of the target company and acquire business intelligence. Traditional approaches for company profiling rely heavily on the availability of rich finance information about the company, such as finance reports and SEC filings, which may not be readily available for many private companies. However, the rapid prevalence of online employment services enables a new paradigm — to obtain the variety of company's information from their employees' online ratings and comments. This, in turn, raises the challenge to develop company profiles from an employee's perspective. To this end, in this paper, we propose a method named Company Profiling based Collaborative Topic Regression (CPCTR), for learning the latent structural patterns of companies. By formulating a joint optimization framework, CPCTR has the ability in collaboratively modeling both textual (e.g., reviews) and numerical information (e.g., salaries and ratings). Indeed, with the identified patterns, including the positive/negative opinions and the latent variable that influences salary, we can effectively carry out opinion analysis and salary prediction. Extensive experiments were conducted on a real-world data set to validate the effectiveness of CPCTR. The results show that our method provides a comprehensive understanding of company characteristics and delivers a more effective prediction of salaries than other baselines.

Customer volume prediction, which predicts the volume from a customer source to a service place, is a very important technique for location selection, market investigation, and other related applications. Most of traditional methods only make use of partial information for either supervised or unsupervised modeling, which cannot well integrate overall available knowledge. In this paper, we propose a method titled GR-NMF for jointly modeling both implicit correlations hidden inside customer volumes and explicit geographical knowledge via an integrated probabilistic framework. The effectiveness of GR-NMF in coupling all-round knowledge is verified over a real-life outpatient dataset under different scenarios. GR-NMF shows particularly evident advantages to all baselines in location selection with the cold-start challenge.

Recent years have witnessed the continuous growth of megalopolises worldwide, which makes urban safety a top priority in modern city life. Among various threats, dangerous goods such as gas and hazardous chemicals transported through and around cities have increasingly become the deadly "bomb" we sleep with every day. In both academia and government, tremendous efforts have been dedicated to dealing with dangerous goods transportation (DGT) issues, but further study is still in great need to quantify the problem and explore its intrinsic dynamics in a big data perspective. In this paper, we present a novel system called DGeye, which features a "duet" between DGT trajectory data and human mobility data for risky zones identification. Moreover, DGeye innovatively takes risky patterns as the keystones in DGT management, and builds causality networks among them for pain point identification, attribution and prediction. Experiments on both Beijing and Tianjin cities demonstrate the effectiveness of DGeye. In particular, the report generated by DGeye driven the Beijing government to lay down gas pipelines for the famous Guijie food street.

Consensus clustering aims to find a single partition of data that agrees as much as possible with existing basic partitions. Given its robustness and generalizability, consensus clustering has emerged as a promising solution to find cluster structures inside heterogeneous big data rising from various application domains. In the area of fuzzy systems, however, research along this line is still in its initial stage with some unsystematic algorithmic studies. Finding a fuzzy consensus partition from multiple fuzzy basic partitions in an efficient, flexible, and robust way is still an exciting open problem calling for further investigation. In light of this, this paper provides a systematic study of fuzzy consensus clustering (FCC) from a utility perspective. Specifically, we first define the objective function of FCC clearly using the novel fuzzified contingency matrix. We then derive a family of FCC Utility functions termed as FCCU that can transform FCC to a weighted piecewise fuzzy $c$ -means clustering (piFCM) problem. This helps us to establish an algorithmic framework for FCC with flexible choice of utility functions, and speeds FCC significantly with a FCM-like iterative process of piFCM. To meet the big data challenge, we further parallelize FCC on the Spark platform with both vertical and horizontal segmentation schemes. Extensive experiments on various real-world datasets demonstrate the excellent performance of FCC, even with a majority of poor basic partitions. In particular, our method exhibits interesting potential for big data clustering in two real-life applications concerned with online event detection and overlapping community detection, respectively.

As a promising way for heterogeneous data analytics, consensus clustering has attracted increasing attention in recent decades. Among various excellent solutions, the co-association matrix based methods form a landmark, which redefines consensus clustering as a graph partition problem. Nevertheless, the relatively high time and space complexities preclude it from wide real-life applications. We, therefore, propose Spectral Ensemble Clustering (SEC) to leverage the advantages of co-association matrix in information integration but run more efficiently. We disclose the theoretical equivalence between SEC and weighted K-means clustering, which dramatically reduces the algorithmic complexity. We also derive the latent consensus function of SEC, which to our best knowledge is the first to bridge co-association matrix based methods to the methods with explicit global objective functions. Further, we prove in theory that SEC holds the robustness, generalizability, and convergence properties. We finally extend SEC to meet the challenge arising from incomplete basic partitions, based on which a row-segmentation scheme for big data clustering is proposed. Experiments on various real-world data sets in both ensemble and multi-view clustering scenarios demonstrate the superiority of SEC to some state-of-the-art methods. In particular, SEC seems to be a promising candidate for big data clustering.

Traffic speed prediction is a long-standing and critically important topic in the area of Intelligent Transportation Systems (ITS). Recent years have witnessed the encouraging potentials of deep neural networks for real-life applications of various domains. Traffic speed prediction, however, is still in its initial stage without making full use of spatio-temporal traffic information. In light of this, in this paper, we propose a deep learning method with an Error-feedback Recurrent Convolutional Neural Network structure (eRCNN) for continuous traffic speed prediction. By integrating the spatio-temporal traffic speeds of contiguous road segments as an input matrix, eRCNN explicitly leverages the implicit correlations among nearby segments to improve the predictive accuracy. By further introducing separate error feedback neurons to the recurrent layer, eRCNN learns from prediction errors so as to meet predictive challenges rising from abrupt traffic events such as morning peaks and traffic accidents. Extensive experiments on real-life speed data of taxis running on the 2nd and 3rd ring roads of Beijing city demonstrate the strong predictive power of eRCNN in comparison to some state-of-the-art competitors. The necessity of weight pre-training using a transfer learning notion has also been testified. More interestingly, we design a novel influence function based on the deep learning model, and showcase how to leverage it to recognize the congestion sources of the ring roads in Beijing.

With the rapid development of online social media, the short text has become the prevalent format for information of Internet. Due to the severe data sparsity issue, accurately discovering knowledge behind these short texts remains a critical challenge. Since regular topic models, such as the Latent Dirichlet Allocation (LDA), can not perform well on short texts, many efforts have been put on building different types of probabilistic topic models for short texts. Inducing topics from dense word-word space instead of sparse document-word space becomes an emerging solution for avoiding data sparsity issue, and the representative one is the Word Network Topic Model (WNTM). However, the word-word space building procedure of WNTM often imports much irrelevant information. In light of this, we propose the Robust WNTM (RWNTM), which can filter out unrelated information during the sampling. The experimental results demonstrate that our method can learn more coherent topics and is more accurate in text classification, as compared with WNTM and other state-of-the-arts.

Recent years have witnessed the unprecedented growth of online social media, which empower short texts as the prevalent format for information of Internet. Given the nature of sparsity, however, short text topic modeling remains a critical yet much-watched challenge in both academy and industry. Rich research efforts have been put on building different types of probabilistic topic models for short texts, among which the self aggregation methods without using auxiliary information become an emerging solution for providing informative cross-text word co-occurrences. However, models along this line are still rarely seen, and the representative one Self-Aggregation Topic Model (SATM) is prone to overfitting and computationally expensive. In light of this, in this paper, we propose a novel probabilistic model called Pseudo-document-based Topic Model (PTM) for short text topic modeling. PTM introduces the concept of pseudo document to implicitly aggregate short texts against data sparsity. By modeling the topic distributions of latent pseudo documents rather than short texts, PTM is expected to gain excellent performance in both accuracy and efficiency. A Sparsity-enhanced PTM (SPTM for short) is also proposed by applying Spike and Slab prior, with the purpose of eliminating undesired correlations between pseudo documents and latent topics. Extensive experiments on various real-world data sets with state-of-the-art baselines demonstrate the high quality of topics learned by PTM and its robustness with reduced training samples. It is also interesting to show that i) SPTM gains a clear edge over PTM when the number of pseudo documents is relatively small, and ii) the constraint that a short text belongs to only one pseudo document is critically important for the success of PTM. We finally take an in-depth semantic analysis to unveil directly the fabulous function of pseudo documents in finding cross-text word co-occurrences for topic modeling.

In this paper, we present multinomial latent logistic regression (MLLR), a new learning paradigm that introduces latent variables to logistic regression. By inheriting the advantages of logistic regression, MLLR is efficiently optimized using the second-order derivatives and provides effective probabilistic analysis on output predictions. MLLR is particularly effective in weakly supervised settings where the latent variable has an exponential number of possible values. The effectiveness of MLLR is demonstrated on four different image understanding applications, including a new challenging architectural style classification task. Furthermore, we show that MLLR can be generalized to general structured output prediction, and in doing so, we provide a thorough investigation of the connections and differences between MLLR and existing related algorithms, including latent structural SVMs and hidden conditional random fields.

Advances in smartphone technology have promoted the rapid development of mobile apps. However, the availability of a huge number of mobile apps in application stores has imposed the challenge of finding the right apps to meet the user needs. Indeed, there is a critical demand for personalized app recommendations. Along this line, there are opportunities and challenges posed by two unique characteristics of mobile apps. First, app markets have organized apps in a hierarchical taxonomy. Second, apps with similar functionalities are competing with each other. Although there are a variety of approaches for mobile app recommendations, these approaches do not have a focus on dealing with these opportunities and challenges. To this end, in this article, we provide a systematic study for addressing these challenges. Specifically, we develop a structural user choice model (SUCM) to learn fine-grained user preferences by exploiting the hierarchical taxonomy of apps as well as the competitive relationships among apps. Moreover, we design an efficient learning algorithm to estimate the parameters for the SUCM model. Finally, we perform extensive experiments on a large app adoption dataset collected from Google Play. The results show that SUCM consistently outperforms state-of-the-art Top-N recommendation methods by a significant margin.

Dynamic topic models (DTM) are of great use toanalyze the evolution of unobserved topics of a text collectionover time. Recent years have witnessed the explosive growth ofstreaming text data emerging from online media, which createsan unprecedented need for DTMs for timely event analysis. While there have been some matrix factorization methods inthe literature for dynamic topic modeling, further study is stillin great need to model emerging, evolving and fading topicsin a more natural and effective way. In light of this, we firstpropose a matrix factorization model called SONMFSR (SoftOrthogonal NMF with Sparse Representation), which makes fulluse of soft orthogonal and sparsity constraints for static topicmodeling. Furthermore, by introducing the constraints of emerging, evolving and fading topics to SONMFSR, we easily obtain a novel DTM called SONMFSRd for dynamic event analysis. Extensive experiments on two public corpora demonstrate the superiority of SONMFSRd to some state-of-the-art DTMs in both topic detection and tracking. In particular, SONMFSRd shows great potential in real-world applications, where popular topics in Two Sessions 2015 are captured and traced dynamically for possible insights.

Driven by profits, spam reviews for product promotion or suppression become increasingly rampant in online shopping platforms. This paper focuses on detecting hidden spam users based on product reviews. In the literature, there have been tremendous studies suggesting diversified methods for spammer detection, but whether these methods can be combined effectively for higher performance remains unclear. Along this line, a hybrid PU-learning-based Spammer Detection (hPSD) model is proposed in this paper. On one hand, hPSD can detect multi-type spammers by injecting or recognizing only a small portion of positive samples, which meets particularly real-world application scenarios. More importantly, hPSD can leverage both user features and user relations to build a spammer classifier via a semi-supervised hybrid learning framework. Experimental results on movie data sets with shilling injection show that hPSD outperforms several state-of-the-art baseline methods. In particular, hPSD shows great potential in detecting hidden spammers as well as their underlying employers from a real-life Amazon data set. These demonstrate the effectiveness and practical value of hPSD for real-life applications.

Aspect-based opinion mining is to find elaborate opinions towards an underlying theme, perspective or viewpoint as to a subject such as a product or an event. Nowadays, with rapid growing of opinionated text on the Web, mining aspect-level opinions has become a promising means for online public opinion analysis. In particular, the booming of various types of online media provide diverse yet complementary information, bringing unprecedented opportunities for public opinion analysis across different populations. Along this line, in this paper, we propose CAMEL, a novel topic model for complementary aspect-based opinion mining across asymmetric collections. CAMEL gains complementarity by modeling both common and specific aspects across different collections, and keeping all the corresponding opinions for contrastive study. To further boost CAMEL, we propose AME, an automatic labeling scheme for maximum entropy model, to help discriminate aspect and opinion words without heavy human labeling. Extensive experiments on synthetic multicollection data sets demonstrate the superiority of CAMEL to baseline methods, in leveraging cross-collection complementarity to find higher-quality aspects and more coherent opinions as well as aspect-opinion relationships. This is particularly true when the collections get seriously imbalanced. Experimental results also show that the AME model indeed outperforms manual labeling in suggesting true opinion words. Finally, case study on two public events further demonstrates the practical value of CAMEL for real-world public opinion analysis.

Ensemble clustering, also known as consensus clustering, is emerging as a promising solution for multi-source and/or heterogeneous data clustering. The co-association matrix based method, which redefines the ensemble clustering problem as a classical graph partition problem, is a landmark method in this area. Nevertheless, the relatively high time and space complexity preclude it from real-life large-scale data clustering. We therefore propose SEC, an efficient Spectral Ensemble Clustering method based on co-association matrix. We show that SEC has theoretical equivalence to weighted K-means clustering and results in vastly reduced algorithmic complexity. We then derive the latent consensus function of SEC, which to our best knowledge is among the first to bridge co-association matrix based method to the methods with explicit object functions. The robustness and generalizability of SEC are then investigated to prove the superiority of SEC in theory. We finally extend SEC to meet the challenge rising from incomplete basic partitions, based on which a scheme for big data clustering can be formed. Experimental results on various real-world data sets demonstrate that SEC is an effective and efficient competitor to some state-of-the-art ensemble clustering methods and is also suitable for big data clustering.

Upon extensive studies, text clustering remains a critical challenge in data mining community. Even by various techniques proposed to overcome some of these challenges, there still exist problems when dealing with weakly related or even noisy features. In response to this, we propose a DIssemble-ASsemble (DIAS) framework for text clustering. DIAS employs simple random feature sampling to disassemble high-dimensional text data and gains diverse structural knowledge. This also does good to avoiding the bulk of noisy features. Then the multi-view knowledge is assembled by weighted Information-theoretic Consensus Clustering (ICC) in order to gain a high-quality consensus partitioning. Extensive experiments on eight real-world text data sets demonstrate the advantages of DIAS over other widely used methods. In particular, DIAS shows strengths in learning from very weak basic partitionings. In addition, it is the natural suitability to distributed computing that makes DIAS become a promising candidate for big text clustering.

The objective of consensus clustering is to find a single partitioning which agrees as much as possible with existing basic partitionings. Consensus clustering emerges as a promising solution to find cluster structures from heterogeneous data. As an efficient approach for consensus clustering, the K-means based method has garnered attention in the literature, however the existing research efforts are still preliminary and fragmented. To that end, in this paper, we provide a systematic study of K-means-based consensus clustering (KCC). Specifically, we first reveal a necessary and sufficient condition for utility functions which work for KCC. This helps to establish a unified framework for KCC on both complete and incomplete data sets. Also, we investigate some important factors, such as the quality and diversity of basic partitionings, which may affect the performances of KCC. Experimental results on various realworld data sets demonstrate that KCC is highly efficient and is comparable to the state-of-the-art methods in terms of clustering quality. In addition, KCC shows high robustness to incomplete basic partitionings with many missing values.

Optimal planning for public transportation is one of the keys to sustainable development and better quality of life in urban areas. Compared to private transportation, public transportation uses road space more efficiently and produces fewer accidents and emissions. In this paper, we focus on the identification and optimization of flawed bus routes to improve utilization efficiency of public transportation services, according to people's real demand for public transportation. To this end, we first provide an integrated mobility pattern analysis between the location traces of taxicabs and the mobility records in bus transactions. Based on mobility patterns, we propose a localized transportation mode choice model, with which we can accurately predict the bus travel demand for different bus routing. This model is then used for bus routing optimization which aims to convert as many people from private transportation to public transportation as possible given budget constraints on the bus route modification. We also leverage the model to identify region pairs with flawed bus routes, which are effectively optimized using our approach. To validate the effectiveness of the proposed methods, extensive studies are performed on real world data collected in Beijing which contains 19 million taxi trips and 10 million bus trips.

Community detection is a long-standing yet very difficult task in social network analysis. It becomes more challenging as many online social networking sites are evolving into super-large scales. Numerous methods have been proposed for community detection from massive networks, but how to reconcile the partitioning efficiency and the community quality remains an open problem. In this paper, we attempt to address this challenge by introducing a COSine-pattern-based COMmunity extraction framework: COSCOM. The COSCOM adopts an extracting view of community detection. It first extracts the so-called asymptotically equivalent structures (AESs) from networks, from which the nodes are further partitioned into crisp communities using any of the existing methods. Specifically, we prove that an AES is a very tight group of nodes, and is actually a cosine pattern defined by the extended cosine similarity. A novel cosine-pattern mining algorithm based on the ordered anti-monotone of cosine similarity is thus proposed for the efficient extraction of AESs. Experiments on various real-world social networks demonstrate the advantage of the extracting view of community detection. In particular, COSCOM shows merits in detecting genuine communities by either internal or external validity.

Recent years have witnessed the explosive growth of online social media. Weibo, a famous "Chinese Twitter", has attracted over half billion users in less than four years. Among them are zombie users or bogus users, who are seemingly active common users but actually marionettes manipulated by intelligent software for economic interests. To probe such users thus becomes critically important for a healthy Weibo, but the existing studies along this line are still in initial stage due to the serious lack of labeled zombies and the limited attributes for user profiling. In light of this, in this paper, we figure out a commercial way for training set labeling, and propose a two-stage cascading model called ProZombie for zombie user recognition. ProZombie decomposes the training/predicting process into fast and refined phases in cascade, which greatly improves the modeling efficiency without sacrificing the accuracy. Moreover, 35 attributes including 16 newly proposed ones are employed for a panoramic description of Weibo users. Experiments on real-world labeled Weibo users demonstrate the effectiveness and efficiency of ProZombie. More interestingly, two case studies based on ProZombie successfully unveil the zombies hidden around common users, and their impact to information propagation on Weibo. To our best knowledge, this study is among the first to quantify these interesting observations on Weibo.

Recent years have witnessed the explosive growth of online social media. Weibo, a famous "Chinese Twitter", has attracted over 0.5 billion users in less than four years, with more than 1000 tweets generated in every second. These tweets are informative but very fragmented, and thus would be better archived from an event perspective, as done by Weibo itself in the "Micro-Topic" program. This effort, however, is yet far from satisfaction for not providing enough analytical power to events. In light of this, in this demo paper, we propose SEA, a System for Event Analysis on Chinese tweets. In general, SEA is an event-centric, multi-functional platform that conducts panoramic analysis on Weibo events from various aspects, including the semantic information of the events, the temporal and spatial trends, the public sentiments, the hidden sub-events, the key users in the event diffusion and their preferences, etc. These functions are enabled by the integration of various analytical models and by the noSQL techniques adopted purposefully for massive tweets management. Finally, a case study on the "Spring Festival" event demonstrates the effectiveness of SEA. To our best knowledge, SEA is the first third-party system that provides panoramic analysis to Weibo events.

Named entity recognition seeks to locate atomic elements in texts and classify them into predefined categories. It is essentially useful for many applications, including microblog analysis and query suggestion. In recent years, with the explosion of Web 2.0, people have found it a promising way to extract large-scale, high-quality entities from structured web content. However, existing studies seldom provide an integrated system for simultaneously extracting and categorizing both the head and tail entities, and the identification of ambiguous entities is still a challenging task. In light of these, we propose a system named quasi-Automatic Named Entity Extraction and Categorization (ANEEC) for massive named-entity management. Specifically, ANEEC first identifies representative websites by using a small seed-set of entities and the query logs of a search engine, and then extracts high-quality entities from the parallel structures in the webpages. ANEEC then employs the extracted entities and their corresponding atom-level groups to establish an entity taxonomy as well as a hierarchical classifier ensemble. Two problems, i.e. definition abnormality and granularity unfitness, have also been addressed to further improve the quality of the taxonomy. An application case using 932 seed entities and the query logs of the search engine Bing demonstrates that ANEEC can effectively identify over 870 000 named entities in 32 bottom-level categories, and the resulting taxonomy has an excellent classification performance with F1=85.17%, provided that the entity features are properly preprocessed and weighted. In particular, ANEEC shows the potential for tail entity recognition and ambiguous entity detection.

Information-theoretic clustering aims to exploit information-theoretic measures as the clustering criteria. A common practice on this topic is the so-called Info-Kmeans, which performs K-means clustering with KL-divergence as the proximity function. While expert efforts on Info-Kmeans have shown promising results, a remaining challenge is to deal with high-dimensional sparse data such as text corpora. Indeed, it is possible that the centroids contain many zero-value features for high-dimensional text vectors, which leads to infinite KL-divergence values and creates a dilemma in assigning objects to centroids during the iteration process of Info-Kmeans. To meet this challenge, in this paper, we propose a Summation-bAsed Incremental Learning (SAIL) algorithm for Info-Kmeans clustering. Specifically, by using an equivalent objective function, SAIL replaces the computation of KL-divergence by the incremental computation of Shannon entropy. This can avoid the zero-feature dilemma caused by the use of KL-divergence. To improve the clustering quality, we further introduce the variable neighborhood search scheme and propose the V-SAIL algorithm, which is then accelerated by a multithreaded scheme in PV-SAIL. Our experimental results on various real-world text collections have shown that, with SAIL as a booster, the clustering performance of Info-Kmeans can be significantly improved. Also, V-SAIL and PV-SAIL indeed help improve the clustering quality at a lower cost of computation.

Clustering validation has long been recognized as one of the vital issues essential to the success of clustering applications. In general, clustering validation can be categorized into two classes, external clustering validation and internal clustering validation. In this paper, we focus on internal clustering validation and present a study of 11 widely used internal clustering validation measures for crisp clustering. The results of this study indicate that these existing measures have certain limitations in different application scenarios. As an alternative choice, we propose a new internal clustering validation measure, named clustering validation index based on nearest neighbors (CVNN), which is based on the notion of nearest neighbors. This measure can dynamically select multiple objects as representatives for different clusters in different situations. Experimental results show that CVNN outperforms the existing measures on both synthetic data and real-world data in different application scenarios.

Shilling attackers apply biased rating profiles to recommender systems for manipulating online product recommendations. Although many studies have been devoted to shilling attack detection, few of them can handle the hybrid shilling attacks that usually happen in practice, and the studies for real-life applications are rarely seen. Moreover, little attention has yet been paid to modeling both labeled and unlabeled user profiles, although there are often a few labeled but numerous unlabeled users available in practice. This paper presents a Hybrid Shilling Attack Detector, or HySAD for short, to tackle these problems. In particular, HySAD introduces MC-Relief to select effective detection metrics, and Semi-supervised Naive Bayes (SNB_lambda) to precisely separate Random-Filler model attackers and Average-Filler model attackers from normal users. Thorough experiments on MovieLens and Netflix datasets demonstrate the effectiveness of HySAD in detecting hybrid shilling attacks, and its robustness for various obfuscated strategies. A real-life case study on product reviews of Amazon.cn is also provided, which further demonstrates that HySAD can effectively improve the accuracy of a collaborative-filtering based recommender system, and provide interesting opportunities for in-depth analysis of attacker behaviors. These, in turn, justify the value of HySAD for real-world applications.

Recent years have witnessed the explosive growth of online social media. Weibo, a Twitter-like online social network in China, has attracted more than 300 million users in less than three years, with more than 1000 tweets generated in every second. These tweets not only convey the factual information, but also reflect the emotional states of the authors, which are very important for understanding user behaviors. However, a tweet in Weibo is extremely short and the words it contains evolve extraordinarily fast. Moreover, the Chinese corpus of sentiments is still very small, which prevents the conventional keyword-based methods from being used. In light of this, we build a system called MoodLens, which to our best knowledge is the first system for sentiment analysis of Chinese tweets in Weibo. In MoodLens, 95 emoticons are mapped into four categories of sentiments, i.e. angry, disgusting, joyful, and sad, which serve as the class labels of tweets. We then collect over 3.5 million labeled tweets as the corpus and train a fast Naive Bayes classifier, with an empirical precision of 64.3%. MoodLens also implements an incremental learning method to tackle the problem of the sentiment shift and the generation of new words. Using MoodLens for real-time tweets obtained from Weibo, several interesting temporal and spatial patterns are observed. Also, sentiment variations are well captured by MoodLens to effectively detect abnormal events in China. Finally, by using the highly efficient Naive Bayes classifier, MoodLens is capable of online real-time sentiment monitoring. The demo of MoodLens can be found at http://goo.gl/8DQ65.

Fuzzyc-means (FCM) is a widely used fuzzy clustering method, which allows an object to belong to two or more clusters with a membership grade between zero and one. Despite the considerable efforts made by the clustering community, the common characteristics of distance functions suitable for FCM remain unclear. To fill this crucial void, in this paper, we first provide a generalized definition of distance functions that fit FCM directly. The goal is to provide more flexibility to FCM in the choice of distance functions while preserving the simplicity of FCM by using the centroids of arithmetic means. Indeed, we show that any distance function that fits FCM directly can be derived by a continuously differentiable convex function and, thus, is an instance of the generalized point-to-centroid distance (P2C-D) by definition. In addition, we prove that if the membership grade matrix is nondegenerate, any instance of the P2C-D fits FCM directly. Finally, extensive experiments have been conducted to demonstrate that the P2C-D leads to the global convergence of FCM and that the clustering performances are significantly affected by the choices of distance functions.

This paper presents a unified view of interestingness measures for interesting pattern discovery. Specifically, we first provide three necessary conditions for interestingness measures being used for association pattern discovery. Then, we reveal one desirable property for interestingness measures: the support-ascending conditional antimonotone property (SA-CAMP). Along this line, we prove that the measures possessing SA-CAMP are suitable for pattern discovery if the itemset-traversal structure is defined by a support-ascending set enumeration tree. In addition, we provide a thorough study on the family of the generalized mean (GM) measure and show their appealing properties, which are exploited for developing the GMiner algorithm for finding interesting association patterns. Finally, experimental results show that GMiner can efficiently identify interesting patterns based on SA-CAMP of the GM measure, even at an extremely low level of support.

Clustering validation has long been recognized as one of the vital issues essential to the success of clustering applications. In general, clustering validation can be categorized into two classes, external clustering validation and internal clustering validation. In this paper, we focus on internal clustering validation and present a detailed study of 11 widely used internal clustering validation measures for crisp clustering. From five conventional aspects of clustering, we investigate their validation properties. Experiment results show that S_Dbw is the only internal validation measure which performs well in all five aspects, while other measures have certain limitations in different application scenarios.

Clustering validation is a long standing challenge in the clustering literature. While many validation measures have been developed for evaluating the performance of clustering algorithms, these measures often provide inconsistent information about the clustering performance and the best suitable measures to use in practice remain unknown. This paper thus fills this crucial void by giving an organized study of 16 external validation measures for K-means clustering. Specifically, we first introduce the importance of measure normalization in the evaluation of the clustering performance on data with imbalanced class distributions. We also provide normalization solutions for several measures. In addition, we summarize the major properties of these external measures. These properties can serve as the guidance for the selection of validation measures in different application scenarios. Finally, we reveal the interrelationships among these external measures. By mathematical transformation, we show that some validation measures are equivalent. Also, some measures have consistent validation performances. Most importantly, we provide a guide line to select the most suitable validation measures for K-means clustering.

This paper studies the generalization and normalization issues of information-theoretic distance measures for clustering validation. Along this line, we first introduce a uniform representation of distance measures, defined as quasi-distance, which is induced based on a general form of conditional entropy. The quasi-distance possesses three properties: symmetry, the triangle law, and the minimum reachable. These properties ensure that the quasi-distance naturally lends itself as the external measure for clustering validation. In addition, we observe that the ranges of the distance measures are different when they apply for clustering validation on different data sets. Therefore, when comparing the performances of clustering algorithms on different data sets, distance normalization is required to equalize ranges of the distance measures. A critical challenge for distance normalization is to obtain the ranges of a distance measure when a data set is provided. To that end, we theoretically analyze the computation of the maximum value of a distance measure for a data set. Finally, we compare the performances of the partition clustering algorithm K-means on various real-world data sets. The experiments show that the normalized distance measures have better performance than the original distance measures when comparing clusterings of different data sets. Also, the normalized Shannon distance has the best performance among four distance measures under study.

K-means is a well-known and widely used partitional clustering method. While there are considerable research efforts to characterize the key features of the K-means clustering algorithm, further investigation is needed to understand how data distributions can have impact on the performance of K-means clustering. To that end, in this paper, we provide a formal and organized study of the effect of skewed data distributions on K-means clustering. Along this line, we first formally illustrate that K-means tends to produce clusters of relatively uniform size, even if input data have varied "true" cluster sizes. In addition, we show that some clustering validation measures, such as the entropy measure, may not capture this uniform effect and provide misleading information on the clustering performance. Viewed in this light, we provide the coefficient of variation (CV) as a necessary criterion to validate the clustering results. Our findings reveal that K-means tends to produce clusters in which the variations of cluster sizes, as measured by CV, are in a range of about 0.3-1.0. Specifically, for data sets with large variation in "true" cluster sizes (e.g., CV > 1.0), K-means reduces variation in resultant cluster sizes to less than 1.0. In contrast, for data sets with small variation in "true" cluster sizes (e.g., CV < 0.3), K-means increases variation in resultant cluster sizes to greater than 0.3. In other words, for the earlier two cases, K-means produces the clustering results which are away from the "true" cluster distributions.

Information-theoretic clustering aims to exploit information-theoretic measures as the clustering criteria. A common practice on this topic is the so-called Info-Kmeans, which performs K-means clustering with KL-divergence as the proximity function. While expert efforts on Info-Kmeans have shown promising results, a remaining challenge is to deal with high-dimensional sparse data such as text corpora. Indeed, it is possible that the centroids contain many zero-value features for high-dimensional text vectors, which leads to infinite KL-divergence values and creates a dilemma in assigning objects to centroids during the iteration process of Info-Kmeans. To meet this challenge, in this paper, we propose a Summation-bAsed Incremental Learning (SAIL) algorithm for Info-Kmeans clustering. Specifically, by using an equivalent objective function, SAIL replaces the computation of KL-divergence by the incremental computation of Shannon entropy. This can avoid the zero-feature dilemma caused by the use of KL-divergence. To improve the clustering quality, we further introduce the variable neighborhood search scheme and propose the V-SAIL algorithm, which is then accelerated by a multithreaded scheme in PV-SAIL. Our experimental results on various real-world text collections have shown that, with SAIL as a booster, the clustering performance of Info-Kmeans can be significantly improved. Also, V-SAIL and PV-SAIL indeed help improve the clustering quality at a lower cost of computation.

K-means is a widely used partitional clustering method. A large amount of effort has been made on finding better proximity (distance) functions for k-means. However, the common characteristics of proximity functions remain unknown. To this end, in this paper, we show that all proximity functions that fit k-means clustering can be generalized as k-means distance, which can be derived by a differentiable convex function. A general proof of sufficient and necessary conditions for k-means distance functions is also provided. In addition, we reveal that k-means has a general uniformization effect; that is, k-means tends to produce clusters with relatively balanced cluster sizes. This uniformization effect of k-means exists regardless of proximity functions. Finally, we have conducted extensive experiments on various real-world data sets, and the results show the evidence of the uniformization effect. Also, we observed that external clustering validation measures, such as entropy and variance of information (VI), have difficulty in measuring clustering quality if data have skewed distributions on class sizes.

Given its importance, the problem of predicting rare classes in large-scale multi-labeled data sets has attracted great attentions in the literature. However, the rare-class problem remains a critical challenge, because there is no natural way developed for handling imbalanced class distributions. This paper thus fills this crucial void by developing a method for Classification using lOcal clusterinG (COG). Specifically, for a data set with an imbalanced class distribution, we perform clustering within each large class and produce sub-classes with relatively balanced sizes. Then, we apply traditional supervised learning algorithms, such as Support Vector Machines (SVMs), for classification. Indeed, our experimental results on various real-world data sets show that our method produces significantly higher prediction accuracies on rare classes than state-of-the-art methods. Furthermore, we show that COG can also improve the performance of traditional supervised learning algorithms on data sets with balanced class distributions.

K-means is a widely used partitional clustering method. While there are considerable research efforts to characterize the key features of K-means clustering, further investigation is needed to reveal whether and how the data distributions can have the impact on the performance of K-means clustering. Indeed, in this paper, we revisit the K-means clustering problem by answering three questions. First, how the "true" cluster sizes can make impact on the performance of K-means clustering? Second, is the entropy an algorithm-independent validation measure for K-means clustering? Finally, what is the distribution of the clustering results by K-means? To that end, we first illustrate that K-means tends to generate the clusters with the relatively uniform distribution on the cluster sizes. In addition, we show that the entropy measure, an external clustering validation measure, has the favorite on the clustering algorithms which tend to reduce high variation on the cluster sizes. Finally, our experimental results indicate that K-means tends to produce the clusters in which the variation of the cluster sizes, as measured by the Coefficient of Variation(CV), is in a specific range, approximately from 0.3 to 1.0.