Modern slavery in global supply chains remains a critical concern, prompting governments in countries such as the UK, Australia, and Canada to introduce corporate transparency legislation, including Modern Slavery Acts. However, the growing volume of compliance reports has made manual analysis increasingly impractical. This challenge is compounded by the fact that many government agencies and NGOs responsible for assessing these statements operate in resource-constrained environments, limiting their ability to deploy large-scale language models. To address both issues, we present a novel AI framework for cross-jurisdictional compliance analysis. Our approach integrates four specialised teacher models, each trained using contrastive learning, prompt engineering, pseudo-labelling, and context-enhanced learning, which are distilled into a single lightweight, computationally efficient student model. This design enhances generalisability across legal frameworks while remaining affordable to deploy in low-resource settings. Experimental evaluations demonstrate that our method achieves higher accuracy and efficiency than existing approaches, offering a scalable, accessible solution for regulators and other stakeholders.

Named Entity Recognition (NER) from open-source security reports has become a crucial task in Cyber Threat Intelligence (CTI) to enable knowledge extraction and support proactive cyber defence. However, existing fully supervised NER approaches struggle to adapt to the dynamic and linguistically complex nature of CTI text. To address these challenges, we propose Label-Efficient Named Entity Recognition (LETNER), a model designed to handle multi-token, sparsely distributed, and fine-grained CTI entity patterns while maintaining low annotation demand. LETNER leverages Convolutional Neural Networks (CNNs) and a gating mechanism to learn dynamic span-based representations, and introduces an orthogonal regularisation to align entity-span relationships in a shared metric space for effective similarity-based inference. Furthermore, a cost-aware evaluation framework is presented to jointly quantify annotation effort and model performance, providing practical insights for decision-making in low-resource settings. Experimental results on a complex CTI dataset containing 22 fine-grained entity classes show that LETNER significantly outperforms baseline models, achieving high performance using only 10% of the annotated data.

Semi-supervised text classification has garnered significant attention for its ability to leverage unlabelled data in settings with limited labelled data. While current state-of-the-art methods employ consistency regularisation with pseudo-labels and augmentation techniques such as synonym replacement and back-translation, they often suffer from high computational cost and limited augmentation diversity. We propose SemiCR-VAE, a novel framework that uses a variational autoencoder with uncertainty-based sampling to generate diverse, task-aware augmentations from the latent space. Experiments on five datasets, including the newly introduced Hydrogen Energy dataset, show that SemiCR-VAE achieves the highest average accuracy (74%) and F1 score (76%), outperforming strong baselines by an average of 3% in accuracy and 4% in F1 score. Our approach exhibits robustness with tuned hyperparameters ( λ s = 100 , λ r l = 0.001 , λ k l = 10 , λ u = 0.5 , T = 0.7 ), ensuring effective text classification in low-resource settings with limited labelled data.

Multimodal learning aims to combine multiple modalities into a joint representation space. Existing works primarily focus on designing multimodal architectures to capture task-specific multimodal features. However, these methods often overlook variations in informativeness across modalities, leading to the feature-level bias problem, where irrelevant information introduces noise into the joint representation space. To address this issue, we propose a novel Deep Kernel Learning enhanced Multimodal Classification (DKLMC) framework. The framework comprises a main network for multimodal classification and a deep kernel learning network for estimating the unimodal contribution for each instance. The DKLMC framework is capable of estimating the contribution of each modality for each instance based on its unimodal contribution factor. This factor is derived using an auxiliary network, and each modality's informativeness is quantified by incorporating the uncertainty of the prediction confidence. Specifically, it addresses the feature-level bias problem by emphasising the more informative unimodal feature and reducing the impact of less informative features in the joint representation space. It features two key innovations: (1) an auxiliary network based on deep kernel learning to estimate unimodal contribution factor effectively, and (2) the use of the Reptile algorithm, a gradient-based meta-learning approach that aims to improve adaptability across various tasks, to balance the optimisation between main and auxiliary networks, ensuring effective learning of both unimodal contribution estimation and multimodal classification tasks. Extensive experiments across diverse datasets demonstrate the effectiveness of the proposed DKLMC framework.

Modern Slavery Acts mandate that corporations disclose their efforts to combat modern slavery, aiming to enhance transparency and strengthen practices for its eradication. However, verifying these statements remains challenging due to their complex, diversified language and the sheer number of statements that must be reviewed. The development of NLP tools to assist in this task is also difficult due to a scarcity of annotated data. Furthermore, as modern slavery transparency legislation has been introduced in several countries, the generalizability of such tools across legal jurisdictions must be studied. To address these challenges, we work with domain experts to make two key contributions. First, we present AIMS.uk and AIMS.ca, newly annotated datasets from the UK and Canada to enable cross-jurisdictional evaluation. Second, we introduce AIMSCheck, an end-to-end framework for compliance validation. AIMSCheck decomposes the compliance assessment task into three levels, enhancing interpretability and practical applicability. Our experiments show that models trained on an Australian dataset generalize well across UK and Canadian jurisdictions, demonstrating the potential for broader application in compliance monitoring. We release the benchmark datasets and AIMSCheck to the public to advance AI-adoption in compliance assessment and drive further research in this field.

Multimodal learning for classification tasks has recently gained significant attention in bioinformatics. Current approaches primarily concentrate on devising efficient deep learning architectures to capture features within and across modalities. However, they typically assume that each modality contributes equally to the classification objective, overlooking inherent biases within multimodal learning. This paper presents a modality contribution confidence-enhanced deep learning framework to address this issue, resulting in an improved fusion space and improved classification performance on multiomics data. Specifically, we propose utilising a non-parametric Gaussian Process to assess the unimodal confidence of each modality and learn within-modality features. Additionally, we introduce the use of the Kullback-Leibler divergence to align multiple modalities and learn cross-modality features. Extensive experiments on four multiomics datasets, incorporating modalities such as static information, DNA, mRNA, miRNA, and protein data, validate the effectiveness of the proposed method. Furthermore, a case study on the blister recovery task is included to demonstrate the practical utility of our model.

Abstract The world of data is vast and complex, harbouring valuable insights and patterns that can drive decision-making in various fields. However, accessing useful information from raw data (e.g. mining project reports) can be a formidable task, often requiring specialised skills and tools. The emergence of generative artificial intelligence (AI) has opened up an intriguing and novel means of engaging with data conversation. This article delves into the novel concept of chatting with organisational data using generative AI. We present an innovative solution that combines a generative AI chatbot (e.g. ChatGPT-QAM) with a language model (e.g. BERT) based extractive question-answer model (BERT-QAM) to generate responses based on given contexts. We use an answer verification model to resolve any disagreements between the responses of ChatGPT-QAM and BERT-QAM. We use the context filtering model to enhance responses by considering valid contexts. Our solution is tested on mining project reports made available by the Geological Survey of Queensland. Through this case study, we highlight several challenges that must be addressed to utilise this approach effectively. The case study shows that the concept of chatting with organisational data can revolutionise how we interact with complex scientific reports, which contain a mix of tables, text and images, to find valuable insights.

• This paper proposes Latent Space Refinement (LSR), a novel unsupervised classification framework that integrates metric learning with clustering-based representation refinement, addressing the critical challenge of label scarcity in cyber threat intelligence (CTI). • LSR introduces a posterior regularisation strategy that aligns latent representations from Pretrained Language Models (PLMs) with an auxiliary TF-IDF-based distribution. This guides unsupervised adaptation to the target domain without any PLM fine-tuning, ensuring scalability and efficiency. • Extensive experiments on three CTI benchmarks demonstrate that LSR consistently outperforms state-of-the-art unsupervised and few-shot baselines in Accuracy and F1 score. • By enabling lightweight unsupervised domain adaptation, LSR offers a plug-and-play solution applicable to CTI, including other resource-constrained domains such as health and legal text classification. Text classification plays a critical role in Cyber Threat Intelligence (CTI) applications, where open-source text data is mined to identify patterns such as Indicators of Compromise (IoC), Tactics, Techniques and Procedures (TTPs), Named Entities and more. However, the dynamic nature of CTI makes traditional supervised machine learning classifiers impractical due to their reliance on large number of labelled training datasets. To address this, we propose Latent Space Refinement (LSR), an unsupervised method designed for CTI text classification. LSR introduces a posterior regularisation strategy where an auxiliary distribution derived from a TF-IDF feature space serves as signals to refine latent representations derrived from Pretrained Language Models (PLMs). By iteratively refining this latent space with clustering signals, LSR enables efficient similarity-based classification using only a few user-provided seed keywords. Extensive experiments on diverse CTI tasks, including both binary and multi-class classification, demonstrate that LSR consistently outperforms state-of-the-art unsupervised and zero-shot/few-shot methods in Accuracy and Weighted F1 score, all without tuning internal PLM parameters. This makes LSR a lightweight and PLM-agnostic solution for real-world CTI applications.

Multi-modal data fusion for effective feature representation in machine learning is challenging due to intrinsic biases present within and across different modalities. Existing multi-modal data fusion methods often face difficulties in learning generic features due to diverse noise patterns and variations in feature dynamics across different modalities. In this paper, we present a novel method called Uncertainty-guided Meta-Learning Multi-modal Fusion and Classification (UMLMC) to address these challenges. UMLMC dynamically transforms multi-modal feature spaces at both the pre- and post-fusion levels by incorporating uncertainty estimates from an auxiliary network. Our model is optimized using a meta-learning algorithm to enhance its generalization capabilities. Extensive experiments on multi-modal data from diverse domains, along with comparisons to state-of-the-art methods, demonstrate the effectiveness of UMLMC in improving classification performance. These results confirm that UMLMC, with its innovative uncertainty estimation and meta-learning framework, effectively learns informative intra- and inter-modal features, leading to superior classification outcomes. • An uncertainty-guided meta-fusion method for multi-modal fusion and classification. • Mitigating the impact of feature-level bias at both before and after fusion. • Meta-learning to generate less biased uncertainty estimation at the feature level.

Attributed graph clustering is a fundamental task in network mining, essential for uncovering valuable insights in various applications. However, the heterogeneity of information from structural and attribute spaces poses significant challenges in achieving consistent and meaningful clustering. To address this, we propose Relative Entropy-based Regularized Non-negative Matrix Factorization (RENMF), a novel approach that integrates structural and attribute information through advanced matrix factorization techniques. RENMF employs Symmetric NMF and Projective NMF to extract community membership distributions from the structural and attribute spaces, respectively. By treating these distributions as homogeneous, RENMF preserves distinct, denoised information from both spaces while considering their heterogeneous complementary information. We introduce Relative Entropy (RE) as a novel regularization term to facilitate interaction between these spaces, aiming to maximize consistency between the discovered latent distributions. In this interaction, we leverage the asymmetric property of RE to emphasize attributes as essential complementary information for structural clustering. The RENMF model is solved using a new iterative multiplicative update rule, with convergence theoretically proven. We evaluate RENMF’s effectiveness through extensive experiments on 10 real-world networks, comparing it to 11 state-of-the-art clustering methods. The results demonstrate RENMF’s superiority in ground truth matching and key quality metrics, outperforming existing methods.

Abstract Generative adversarial networks (GANs) have demonstrated their effectiveness in generating temporal data to fill in missing values, enhancing the classification performance of time series data. Longitudinal datasets encompass multivariate time series data with additional static features that contribute to sample variability over time. These datasets often encounter missing values due to factors such as irregular sampling. However, existing GAN-based imputation methods that address this type of data missingness often overlook the impact of static features on temporal observations and classification outcomes. This paper presents a novel method, fusion-aided imputer-classifier GAN (FaIC-GAN), tailored for longitudinal data classification. FaIC-GAN simultaneously leverages partially observed temporal data and static features to enhance imputation and classification learning. We present four multimodal fusion strategies that effectively extract correlated information from both static and temporal modalities. Our extensive experiments reveal that FaIC-GAN successfully exploits partially observed temporal data and static features, resulting in improved classification accuracy compared to unimodal models. Our post-additive and attention-based multimodal fusion approaches within the FaIC-GAN model consistently rank among the top three methods for classification.

Multi-modal representation learning has received significant attention across diverse research domains due to its ability to model a scenario comprehensively. Learning the cross-modal interactions is essential to combining multi-modal data into a joint representation. However, conventional cross-attention mechanisms can produce noisy and non-meaningful values in the absence of useful cross-modal interactions among input features, thereby introducing uncertainty into the feature representation. These factors have the potential to degrade the performance of downstream tasks. This paper introduces a novel Pre-gating and Contextual Attention Gate (PCAG) module for multi-modal learning comprising two gating mechanisms that operate at distinct information processing levels within the deep learning model. The first gate filters out interactions that lack informativeness for the downstream task, while the second gate reduces the uncertainty introduced by the cross-attention module. Experimental results on eight multi-modal classification tasks spanning various domains show that the multi-modal fusion model with PCAG outperforms state-of-the-art multi-modal fusion models. Additionally, we elucidate how PCAG effectively processes cross-modality interactions.

Early classification of longitudinal data remains an active area of research today. The complexity of these datasets and the high rates of missing data caused by irregular sampling present data-level challenges for the Early Longitudinal Data Classification (ELDC) problem. Coupled with the algorithmic challenge of optimising the opposing objectives of early classification (i.e., earliness and accuracy), ELDC becomes a non-trivial task. Inspired by the generative power and utility of the Generative Adversarial Network (GAN), we propose a novel context-conditional, longitudinal early classifier GAN (LEC-GAN). This model utilises informative missingness, static features and earlier observations to improve the ELDC objective. It achieves this by incorporating ELDC as an auxiliary task within an imputation optimization process. Our experiments on several datasets demonstrate that LEC-GAN outperforms all relevant baselines in terms of F1 scores while increasing the earliness of prediction.

Clustering on the data with multiple aspects, such as multi-view or multi-type relational data, has become popular in recent years due to their wide applicability. The approach using manifold learning with the Non-negative Matrix Factorization (NMF) framework, that learns the accurate low-rank representation of the multi-dimensional data, has shown effectiveness. We propose to include the inter-manifold in the NMF framework, utilizing the distance information of data points of different data types (or views) to learn the diverse manifold for data clustering. Empirical analysis reveals that the proposed method can find partial representations of various interrelated types and select useful features during clustering. Results on several datasets demonstrate that the proposed method outperforms the state-of-the-art multi-aspect data clustering methods in both accuracy and efficiency.

Language model (LM) has become a common method of transfer learning in Natural Language Processing (NLP) tasks when working with small labeled datasets. An LM is pretrained using an easily available large unlabelled text corpus and is fine-tuned with the labelled data to apply to the target (i.e., downstream) task. As an LM is designed to capture the linguistic aspects of semantics, it can be biased to linguistic features. We argue that exposing an LM model during fine-tuning to instances that capture diverse semantic aspects (e.g., topical, linguistic, semantic relations) present in the dataset will improve its performance on the underlying task. We propose a Mixed Aspect Sampling (MAS) framework to sample instances that capture different semantic aspects of the dataset and use the ensemble classifier to improve the classification performance. Experimental results show that MAS performs better than random sampling as well as the state-of-the-art active learning models to abuse detection tasks where it is hard to collect the labelled data for building an accurate classifier.

Coupled Matrix Tensor Factorization (CMTF) facilitates the integration and analysis of multiple data sources and helps discover meaningful information. Nonnegative CMTF (N-CMTF) has been employed in many applications for identifying latent patterns, prediction, and recommendation. However, due to the added complexity with coupling between tensor and matrix data, existing N-CMTF algorithms exhibit poor computation efficiency. In this paper, a computationally efficient N-CMTF factorization algorithm is presented based on the column-wise element selection, preventing frequent gradient updates. Theoretical and empirical analyses show that the proposed N-CMTF factorization algorithm is not only more accurate but also more computationally efficient than existing algorithms in approximating the tensor as well as in identifying the underlying nature of factors.

Effective methods are required to be developed that can deal with the multi-faceted nature of the multi-view data. We design a factorization-based loss function-based method to simultaneously learn two components encoding the consensus and complementary information present in multi-view data by using the Coupled Matrix Factorization (CMF) and Non-negative Matrix Factorization (NMF). We propose a novel optimal manifold for multi-view data which is the most consensed manifold embedded in the high-dimensional multi-view data. A new complementary enhancing term is added in the loss function to enhance the complementary information inherent in each view. An extensive experiment with diverse datasets, benchmarking the state-of-the-art multi-view clustering methods, has demonstrated the effectiveness of the proposed method in obtaining accurate clustering solution.

Outlier detection in text data collections has become significant due to the need of finding anomalies in the myriad of text data sources. High feature dimensionality, together with the larger size of these document collections, presents a need for developing accurate outlier detection methods with high efficiency. Traditional outlier detection methods face several challenges including data sparseness, distance concentration, and the presence of a larger number of sub-groups when dealing with text data. In this article, we propose to address these issues by developing novel concepts such as presenting documents with the rare document frequency, finding ranking-based neighborhood for similarity computation, and identifying sub-dense local neighborhoods in high dimensions. To improve the proposed primary method based on rare document frequency, we present several novel ensemble approaches using the ranking concept to reduce the false identifications while finding the higher number of true outliers. Extensive empirical analysis shows that the proposed method and its ensemble variations improve the quality of outlier detection in document repositories as well as they are found scalable compared to the relevant benchmarking methods.

With the advancements in computing technology and web-based applications, data are increasingly generated in multi-dimensional form. These data are usually sparse due to the presence of a large number of users and fewer user interactions. To deal with this, the Nonnegative Tensor Factorization (NTF) based methods have been widely used. However existing factorization algorithms are not suitable to process in all three conditions of size, density, and rank of the tensor. Consequently, their applicability becomes limited. In this article, we propose a novel fast and efficient NTF algorithm using the element selection approach. We calculate the element importance using Lipschitz continuity and propose a saturation point-based element selection method that chooses a set of elements column-wise for updating to solve the optimization problem. Empirical analysis reveals that the proposed algorithm is scalable in terms of tensor size, density, and rank in comparison to the relevant state-of-the-art algorithms.

Knowledge Graphs (KGs) have been proven to be incredibly useful for enriching semantic Web search results and allowing queries with a well-defined result set. In recent years much attention has been given to the task of inferring missing facts based on existing facts in a KG. Approaches have also been proposed for inferring types of entities, however these are successful in common types such as 'Person', 'Movie', or 'Actor'. There is still a large gap, however, in the inference of fine-grained types which are highly important for exploring specific lists and collections within web search. Generally there are also relatively fewer observed instances of fine-grained types present to train in KGs, and this poses challenges for the development of effective approaches. In order to address the issue, this paper proposes a new approach to the fine-grained type inference problem. This new approach is explicitly modeled for leveraging domain knowledge and utilizing additional data outside KG, that improves performance in fine-grained type inference. Further improvements in efficiency are achieved by extending the model to probabilistic inference based on entity similarity and typed class classification. We conduct extensive experiments on type triple classification and entity prediction tasks on Freebase FB15K benchmark dataset. The experiment results show that the proposed model outperforms the state-of-the-art approaches for type inference in KG, and achieves high performance results in many-to-one relation in predicting tail for KG completion task.

In this paper, we present a mobile crowdsensing approach to understand the daily lifestyle of the older population in Singapore. By implementing novel clustering, sensor fusion, and user profiling techniques to analyze the multisensor data (location, noise, and light) collected from a smartphone application, we identified the travel patterns at several points of interest (POI), the impact of travel frequency for certain POI, and three main user profiles. The results show that older adults mostly spend time at food courts and community centers in their home neighborhood, but they travel away from the neighborhood for healthcare and religious purposes. We found that POIs have more visits if they are easily accessible (in terms of travel time from home) regardless of the distance from home.

In this paper, we introduce a novel problem of discovering influence hierarchy to organize influential users in a social network into different levels according to their potential of spreading influence. We present a novel approach of discovering influence hierarchy utilizing the temporal aspect and flow direction of interactions among users. The influence hierarchy has the potential to visualize the information flow of the network and identify different roles such as creators, information disseminators, emerging leaders and active followers. It is highly applicable in several domains such as sociology, marketing, political science and disaster management.

Non-negative Matrix Factorization (NMF) methods have been effectively used for clustering high dimensional data. Manifold learning is combined with the NMF framework to ensure the projected lower dimensional representations preserve the local geometric structure of data. In this paper, considering the context of multi-type relational data clustering, we develop a new formulation of manifold learning to be embedded in the factorization process such that the new low-dimensional space can maintain both local and global structures of original data. We also propose to include the interactions between clusters of different data types by enforcing a Normalize Cut-type constraint that leads to a comprehensive NMF-based framework. A theoretical analysis and extensive experiments are provided to validate the effectiveness of the proposed work.

With the advancements in computing and location-acquisition technologies, large volumes of spatio-temporal tra-jectory data are being generated and stored. Anomaly detection in trajectory data is significant for several applications. Using a data-driven spatio-temporal context in the form of geographical sub-regions and different time-periods can enhance the relevance of detected anomalies. We propose a novel scalable contextual anomaly detection method for trajectory data using the regional density information. The effectiveness and scalability of the proposed method is shown through the empirical analysis and benchmarking with the state-of-the-art method.

This paper proposes a corpus-based media posts expansion technique with a density-based clustering method for community detection. To enrich the user content information, firstly all (short-text) media posts of a user are combined with hash tags and URLs available with the posts. The expanded content view is further augmented by the virtual words inferred using the novel concept of matrix factorization based topic proportion vector approximation. This expansion technique deals with the extreme sparseness of short text data which otherwise leads to insufficient word co-occurrence and, in hence, inaccurate outcome. We then propose to group these augmented posts which represent users by identifying the density patches and form user communities. The remaining isolated users are then assigned to communities to which they are found most similar using a distance measure. Experimental results using several Twitter datasets show that the proposed approach is able to deal with common issues attached with (short-text) media posts to form meaningful communities and attain high accuracy compared to relevant benchmarking methods.

In this paper we introduce the flow hierarchy derived from frequent interactions to represent complex social networks in a concise way. Frequent interactions extract the impactful users while the flow hierarchy visualizes the dependencies and identifies different roles such as leaders, topic experts, information disseminators, emerging leaders and active followers. It is highly applicable in intelligent systems which involve user ranking, expert searching, recommendation, viral marketing, political campaigning, disaster management and many more. We present novel methods of deriving flow hierarchy considering the temporal dimension of interactions among users, flow directions and structural dependencies. We empirically evaluate proposed methods using real-world social network interaction datasets related to citation and retweet networks. Empirical analysis reveals that a hierarchy derived from frequent social interactions is effective in extracting the impactful users and their position in the network. Baseline results with user-centric measures show the efficacy of the proposed methods in finding a concise network representation.

Discriminative itemsets can be more useful than frequent itemsets as the former identifies the frequent itemsets in one dataset with much higher frequencies than the same itemsets in other datasets. The discriminative itemsets can distinguish the target dataset from all others. The discriminative itemsets are a small subset of frequent itemsets. The efficient mining of discriminative itemsets is a challenging problem, since the Apriori property of frequent itemsets is not applicable, and the designed algorithms must deal with the exponential number of itemset combinations in more than one dataset. In this paper, a novel algorithm, called DISSparse, is proposed for efficient mining of discriminative itemsets. Two determinative heuristics are proposed for limiting the mining of discriminative itemsets to the potential discriminative itemsets. Our experiments show the efficient time and space usage of the proposed algorithm in the large and complex datasets.

Clustering can help to make large datasets more manageable by grouping together similar objects. However, most clustering approaches are unable to scale to very large datasets (e.g. more than 10 million objects). The K-Tree is a data structure and clustering algorithm that has proven to be scalable with large streaming datasets. Here, we apply the K-Tree to spatial data (satellite images) and extend from a single threaded to a multicore environment. We show that the K-Tree is able to cluster larger dataset more efficiently than baseline approaches.

Location-aware social media is increasing being used to inform decisions in a spatiotemporal context. However, collecting, fusing, processing and merging information from different social media platforms is a challenge because of diversity of information between different platforms. Here, we present the WIMBY, which is able to access multiple social media platforms to help users answer the question "What's in my Backyard?". In doing so, the WIMBY helps to address the challenge of dealing with diverse social media information. It is believed that the WIMBY can be extended to include more information sources (including other social media platforms) to help inform decision makers in a wider array of applications.

For the task of tag-based item recommendations, the underlying tensor model faces several challenges such as high data sparsity and inferring latent factors effectively. To overcome the inherent sparsity issue of tensor models, we propose the graded-relevance interpretation scheme that leverages the tagging data effectively. Unlike the existing schemes, the graded-relevance scheme interprets the tagging data richly, differentiates the non-observed tagging data insightfully, and annotates each entry as one of the "relevant", "likely relevant", "irrelevant", or "indecisive" labels. To infer the latent factors of tensor models correctly to produce the high quality recommendation, we develop a novel learning-to-rank method, Go-Rank, that optimizes Graded Average Precision (GAP). Evaluating the proposed method on real-world datasets, we show that the proposed interpretation scheme produces a denser tensor model by revealing "relevant" entries from the previously assumed "irrelevant" entries. Optimizing GAP as the ranking metric, the quality of the recommendations generated by Go-Rank is found superior against the benchmarking methods.

High-Order Co-Clustering (HOCC) methods have attracted high attention in recent years because of their ability to cluster multiple types of objects simultaneously using all available information. During the clustering process, HOCC methods exploit object co-occurrence information, i.e., inter-type relationships amongst different types of objects as well as object affinity information, i.e., intra-type relationships amongst the same types of objects. However, it is difficult to learn accurate intra-type relationships in the presence of noise and outliers. Existing HOCC methods consider the p nearest neighbours based on Euclidean distance for the intra-type relationships, which leads to incomplete and inaccurate intra-type relationships. In this paper, we propose a novel HOCC method that incorporates multiple subspace learning with a heterogeneous manifold ensemble to learn complete and accurate intra-type relationships. Multiple subspace learning reconstructs the similarity between any pair of objects that belong to the same subspace. The heterogeneous manifold ensemble is created based on two-types of intra-type relationships learnt using p-nearest-neighbour graph and multiple subspaces learning. Moreover, in order to make sure the robustness of clustering process, we introduce a sparse error matrix into matrix decomposition and develop a novel iterative algorithm. Empirical experiments show that the proposed method achieves improved results over the state-of-art HOCC methods for FScore and NMI.

The proliferation of the web presents an unsolved problem of automatically analyzing billions of pages of natural language. We introduce a scalable algorithm that clusters hundreds of millions of web pages into hundreds of thousands of clusters. It does this on a single mid-range machine using efficient algorithms and compressed document representations. It is applied to two web-scale crawls covering tens of terabytes. ClueWeb09 and ClueWeb12 contain 500 and 733 million web pages and were clustered into 500,000 to 700,000 clusters. To the best of our knowledge, such fine grained clustering has not been previously demonstrated. Previous approaches clustered a sample that limits the maximum number of discoverable clusters. The proposed EM-tree algorithm uses the entire collection in clustering and produces several orders of magnitude more clusters than the existing algorithms. Fine grained clustering is necessary for meaningful clustering in massive collections where the number of distinct topics grows linearly with collection size. These fine-grained clusters show an improved cluster quality when assessed with two novel evaluations using ad hoc search relevance judgments and spam classifications for external validation. These evaluations solve the problem of assessing the quality of clusters where categorical labeling is unavailable and unfeasible.

Detecting anomalies in the online social network is a significant task as it assists in revealing the useful and interesting information about the user behavior on the network. This paper proposes a rule-based hybrid method using graph theory, Fuzzy clustering and Fuzzy rules for modeling user relationships inherent in online-social-network and for identifying anomalies. Fuzzy C-Means clustering is used to cluster the data and Fuzzy inference engine is used to generate rules based on the cluster behavior. The proposed method is able to achieve improved accuracy for identifying anomalies in comparison to existing methods.

Road surface skid resistance has been shown to have a strong relationship to road crash risk, however, applying the current method of using investigatory levels to identify crash prone roads is problematic as they may fail in identifying risky roads outside of the norm. The proposed method analyses a complex and formerly impenetrable volume of data from roads and crashes using data mining. This method rapidly identifies roads with elevated crash-rate, potentially due to skid resistance deficit, for investigation. A hypothetical skid resistance/crash risk curve is developed for each road segment, driven by the model deployed in a novel regression tree extrapolation method. The method potentially solves the problem of missing skid resistance values which occurs during network-wide crash analysis, and allows risk assessment of the major proportion of roads without skid resistance values.

A new community and communication type of social networks - online dating - are gaining momentum. With many people joining in the dating network, users become overwhelmed by choices for an ideal partner. A solution to this problem is providing users with partners recommendation based on their interests and activities. Traditional recommendation methods ignore the users' needs and provide recommendations equally to all users. In this paper, we propose a recommendation approach that employs different recommendation strategies to different groups of members. A segmentation method using the Gaussian Mixture Model (GMM) is proposed to customize users' needs. Then a targeted recommendation strategy is applied to each identified segment. Empirical results show that the proposed approach outperforms several existing recommendation methods.

In a people-to-people matching systems, filtering is widely applied to find the most suitable matches. The results returned are either too many or only a few when the search is generic or specific respectively. The use of a sophisticated recommendation approach becomes necessary. Traditionally, the object of recommendation is the item which is inanimate. In online dating systems, reciprocal recommendation is required to suggest a partner only when the user and the recommended candidate both are satisfied. In this paper, an innovative reciprocal collaborative method is developed based on the idea of similarity and common neighbors, utilizing the information of relevance feedback and feature importance. Extensive experiments are carried out using data gathered from a real online dating service. Compared to benchmarking methods, our results show the proposed method can achieve noticeable better performance.

Grouping users in social networks is an important process that improves matching and recommendation activities in social networks. The data mining methods of clustering can be used in grouping the users in social networks. However, the existing general purpose clustering algorithms perform poorly on the social network data due to the special nature of users' data in social networks. One main reason is the constraints that need to be considered in grouping users in social networks. Another reason is the need of capturing large amount of information about users which imposes computational complexity to an algorithm. In this paper, we propose a scalable and effective constraint-based clustering algorithm based on a global similarity measure that takes into consideration the users' constraints and their importance in social networks. Each constraint's importance is calculated based on the occurrence of this constraint in the dataset. Performance of the algorithm is demonstrated on a dataset obtained from an online dating website using internal and external evaluation measures. Results show that the proposed algorithm is able to increases the accuracy of matching users in social networks by 10% in comparison to other algorithms.

Road networks are a vital national property however, with the passage of time their condition deteriorates. It is critical for a road agency to know the various factors such as road conditions, traffic conditions, environmental conditions that affect the Road Pavement Deflection values. This paper proposes a data analytic application for assessing the road pavement condition. The data analytics process includes acquisition and integration of data from multiple sources, pre-processing the data and mining the useful information from the data. The generated data mining models are able to demonstrate factors that affect pavement deflection data. Outputs of this research inform the road managers about the road conditions and enable them in formulating an efficient e-government policy for budget and maintenance of this road asset.

The rapid growth in the number of users using social networks and the information that a social network requires about their users make the traditional matching systems insufficiently adept at matching users within social networks. This paper introduces the use of clustering to form communities of users and, then, uses these communities to generate matches. Forming communities within a social network helps to reduce the number of users that the matching system needs to consider, and helps to overcome other problems from which social networks suffer, such as the absence of user activities' information about a new user. The proposed system has been evaluated on a dataset obtained from an online dating website. Empirical analysis shows that accuracy of the matching process is increased using the community information.

A new relationship type of social networks - online dating - are gaining popularity. With a large member base, users of a dating network are overloaded with choices about their ideal partners. Recommendation methods can be utilized to overcome this problem. However, traditional recommendation methods do not work effectively for online dating networks where the dataset is sparse and large, and a two-way matching is required. This paper applies social networking concepts to solve the problem of developing a recommendation method for online dating networks. We propose a method by using clustering, SimRank and adapted SimRank algorithms to recommend matching candidates. Empirical results show that the proposed method can achieve nearly double the performance of the traditional collaborative filtering and common neighbor methods of recommendation.

Developing safe and sustainable road systems is a common goal in all countries. Applications to assist with road asset management and crash minimization are sought universally. This paper presents a data mining methodology using decision trees for modeling the crash proneness of road segments using available road and crash attributes. The models quantify the concept of crash proneness and demonstrate that road segments with only a few crashes have more in common with non-crash roads than roads with higher crash counts. This paper also examines ways of dealing with highly unbalanced data sets encountered in the study.

The World Wide Web no longer consists just of HTML pages. Our work sheds light on a number of trends on the Internet that go beyond simple Web pages. The hidden Web provides a wealth of data in semi-structured form, accessible through Web forms and Web services. These services, as well as numerous other applications on the Web, commonly use XML, the eXtensible Markup Language. XML has become the lingua franca of the Internet that allows customized markups to be defined for specific domains. On top of XML, the Semantic Web grows as a common structured data source. In this work, we first explain each of these developments in detail. Using real-world examples from scientific domains of great interest today, we then demonstrate how these new developments can assist the managing, harvesting, and organization of data on the Web. On the way, we also illustrate the current research avenues in these domains. We believe that this effort would help bridge multiple database tracks, thereby attracting researchers with a view to extend database technology.

Item folksonomy or tag information is popularly available on the web now. However, since tags are arbitrary words given by users, they contain a lot of noise such as tag synonyms, semantic ambiguities and personal tags. Such noise brings difficulties to improve the accuracy of item recommendations. In this paper, we propose to combine item taxonomy and folksonomy to reduce the noise of tags and make personalized item recommendations. The experiments conducted on the dataset collected from Amazon.com demonstrated the effectiveness of the proposed approaches. The results suggested that the recommendation accuracy can be further improved if we consider the viewpoints and the vocabularies of both experts and users.

This paper proposes a hybrid approach of personalized Web Information Retrieval that utilizes (1) ontology for retrieval of user's context (2) user profile that is temporarily updated according to users' browsing behavior and (3) collaborative filtering for considering recommendation of similar users. Empirical analysis reveals that Precision, Recall and F-Score of most of the queries for many users are improved with using the proposed method.

This paper introduces a clustering approach, XML Clustering using Frequent Substructures (XCFS) that considers both the structural and the content information of XML documents in clustering. XCFS uses frequent substructures in the form of a novel representation, Closed Frequent Embedded (CFE) subtrees to constrain the content in the clustering process. The empirical analysis ascertains that XCFS can effectively cluster even very large XML datasets and outperforms other existing methods.

This paper proposes a novel Hybrid Clustering approach for XML documents (HCX) that first determines the structural similarity in the form of frequent subtrees and then uses these frequent subtrees to represent the constrained content of the XML documents in order to determine the content similarity. The empirical analysis reveals that the proposed method is scalable and accurate.

The social tags in web 2.0 are becoming another important information source to profile users' interests and preferences to make personalized recommendations. To solve the problem of low information sharing caused by the free-style vocabulary of tags and the long tails of the distribution of tags and items, this paper proposes an approach to integrate the social tags given by users and the item taxonomy with standard vocabulary and hierarchical structure provided by experts to make personalized recommendations. The experimental results show that the proposed approach can effectively improve the information sharing and recommendation accuracy.

Recommender systems' performance can be easily affected when there are no sufficient item preferences data provided by previous users, and it is commonly referred to as cold-start problem. This paper suggests another information source, item taxonomies, in addition to item preference data for assisting recommendation making. Item taxonomy information has been popularly applied in diverse ecommerce domains for product or content classification, and therefore can be easily obtained and adapted by recommender systems. In this paper, we investigate the implicit relations between users' item preferences and taxonomic preferences, suggest and verify that users who share similar item preferences may also share similar taxonomic preferences. Under this assumption, a novel recommendation technique is proposed that combines the users' item preferences and the additional taxonomic preferences together to make better quality recommendations as well as alleviate the cold-start problem. Empirical evaluations to this approach are conducted and the results show that the proposed technique outperforms other existing techniques in both recommendation quality and computation efficiency.

An online question answering (QA) portal provides users a way to socialize and help each other to solve problems. The majority of the online question answer systems use user-feedback to rank userspsila answers. This way of ranking is inefficient as it involves ongoing efforts by the users and is subjective. Currently researchers have utilized link analysis of user interactions for this task. However, this is not accurate in some circumstances. A detailed structural analysis of an online QA portal is conducted in this paper. A novel approach based on userspsila reputation reflecting the usage patterns is proposed to rank and recommend the user answers. The method is compared with a popular link topology analysis method, HITS. The result of the proposed method is promising.

Retrieving accurate information from the Web is a great challenge to users. The existing information retrieval systems are mostly term-based and thus need to be enhanced toward knowledge-based. User information needs need to be better captured in order to deliver personalized search results. In this paper, an ontology-based framework is proposed for capturing user information needs using a world knowledge base and the user's local instance repository. The framework aims to discover a user's background knowledge for knowledge retrieval. The evaluation result is encouraging, in which the proposed model achieved the same performance as a manual user model.

Due to the lack of semantic descriptions of the Web services, the search results returned by the service registries are effectively inadequate. This paper presents the Semantic Web services Clustering (SWSC) method that extends the semantic representation of services and groups the similar Web services in order to improve the service discovery. The empirical analysis shows the improvement in service discovery with the use of SWSC.

It is well accepted that ontology is useful for personalized Web information gathering. However, it is challenging to use semantic relations of "kind-of", "part-of", and "related-to" and synthesize commonsense and expert knowledge in a single computational model. In this paper, a personalized ontology model is proposed attempting to answer this challenge. A two-dimensional (Exhaustivity and Specificity) method is also presented to quantitatively analyze these semantic relations in a single framework. The proposals are successfully evaluated by applying the model to a Web information gathering system. The model is a significant contribution to personalized ontology engineering and concept-based Web information gathering in Web Intelligence.

Since the emergence in the popularity of XML for data representation and exchange over the Web, the distribution of XML documents has rapidly increased. It has become a challenge for researchers to turn these documents into a more useful information utility. In this paper, we introduce a novel clustering algorithm PCXSS that keeps the heterogeneous XML documents into various groups according to their similar structural and semantic representations. We develop a global criterion function CPSim that progressively measures the similarity between a XML document and existing clusters, ignoring the need to compute the similarity between two individual documents. The experimental analysis shows the method to be fast and accurate.

This paper discusses the influence of semantic computation in structural data such as XML for clustering similar data. We study how the semantic similarity at individual element level influences the overall similarity of documents. The empirical results indicate that the semantic measures do not play an important role in finding clusters in structural data such as XML

The traditional techniques rely on human effort to acquire training sets, which is expensive and inefficient. In this paper we present an alternative method to automatically acquire training sets without heavy investment of user efforts. The proposed method tends to fill a gap for effectiveness of using Web data in Web mining, and contributes to Web information gathering. The evaluation shows that the method is adequate to yield an promising achievement.

Most existing recommender systems nowadays operate in a single organizational base, and very often they do not have sufficient resources to be used in order to generate quality recommendations. Therefore, it would be beneficial if recommender systems of different organizations can cooperate together to share their resources and recommendations. In this paper, we present a distributed recommender system model that consists of multiple recommender systems from different organizations. With the hope to provide better recommendation service to users, the recommender systems can improve their performances by sharing their recommendations cooperatively. A recommender selection technique based on the Gittins indices is presented in this paper, and it makes selections based on the stability, average performance and selection frequency of the recommenders

Data mining techniques provide people with new power to research and manipulate the existing large volume of data. A data mining process discovers interesting information from the hidden data that can either be used for future prediction and/or intelligently summarising the details of the data. There are many achievements of applying data mining techniques to various areas such as marketing, medical, and financial, although few of them can be currently seen in software engineering domain. In this paper, a proposed data mining application in software engineering domain is explained and experimented. The empirical results demonstrate the capability of data mining techniques in software engineering domain and the potential benefits in applying data mining to this area.