Accurate traffic flow prediction is fundamental to intelligent transportation systems, yet it remains challenging due to the complex spatio-temporal dynamics of road networks. Most existing approaches model traffic conditions as discrete snapshots, which fail to capture the underlying continuous evolution of traffic states and lead to significant information loss. Moreover, many methods insufficiently model spatial heterogeneity, inadequately distinguishing between local geographic proximity and global semantic correlations. To address these limitations, we propose Spatio-Temporal Forecasting with Adaptive Continuous learning and Heterogeneity awareness (STFACH), a novel framework for traffic forecasting. First, we introduce an Adaptive Continuous Learning module that bridges discrete observations and continuous dynamics, while capturing intrinsic volatility often filtered out by deterministic interpolation. Driven by a learnable control path derived from intrinsic traffic volatility, this module leverages attentive neural differential equations to reconstruct continuous latent trajectories and explicitly model fine-grained temporal evolution. Second, we design a Heterogeneity-Aware module to resolve spatial ambiguity by disentangling spatial dependencies into dual branches, constructing adaptive adjacency matrices that jointly capture local geometric relations and global semantic affinities. Extensive experiments on ten real-world benchmark datasets spanning diverse traffic scenarios demonstrate that STFACH significantly outperforms state-of-the-art baselines, validating its effectiveness and robustness in modeling both continuous dynamics and structural heterogeneity. The code is publicly available at https://github.com/MaxRubby/STFACH_2026.

Accurate identification of pathogens-especially those with pandemic potential-remains a significant challenge, particularly when traditional sequence alignment methods fail. While recent genome sequence identification methods have shown promise, most do not account for the hierarchical structure of biological taxonomy or the varying informativeness of genomic features across classification levels. To address these limitations, we propose HieRMVir (Hierarchical Random forest and Mutual information-based Viral genome classifier), a novel hierarchical deep learning framework that integrates random forest (RF)-based feature weighting with mutual information (MI)-guided attention regularization for interpretable and accurate viral sequence classification. HieRMVir performs classification across three levels and leverages feature importance scores from RF to scale input features, while MI scores are used to guide the attention mechanism towards statistically informative k-mer patterns through regularized loss. Experimental results on over one million genome sequences demonstrate that HieRMVir achieves an average accuracy of 95.8% (95% CI: 95.3-96.4%), outperforming existing methods on multiple metrics. Evaluation using hierarchical performance metrics and the analysis of learned attention weights further highlight the biological relevance and interpretability of HieRMVir.

The rapid expansion of protein sequence databases has created challenges for efficient storage, transmission, and analysis. Unlike genomic sequences with only four nucleotide bases, proteins are composed of twenty amino acids, making compression more complex. Existing specialized protein compressors, such as AC, AC2 and CPM-FCM, have achieved promising performance but still face limitations, including high computational cost, low adaptability and limited biological interpretability. This paper introduces GMP (Genetic algorithm-based MDL Protein compressor), a novel protein compression framework that leverages the Minimum Description Length (MDL) principle with a genetic algorithm to discover optimal patterns of amino acid subsequences (kAA-mers). Experimental results demonstrate that GMP attains compression performance comparable to state-of-the-art methods while additionally supporting tasks such as classification and clustering-capabilities absent from traditional protein compressors. This makes GMP not only an efficient compression framework but also a biologically interpretable tool for protein sequence analysis. GMP is available at github.com/MuhammadzohaibNawaz/GMP

Longitudinal gene expression profiling provides crucial insights into dynamic biological processes, yet predictive modeling is severely hampered by irregular sampling, missing values, and high dimensionality. Mainstream approaches, including recurrent and transformer-based architectures, often struggle with the sparsity and temporal misalignment inherent in multi-visit transcriptomic datasets. We introduce G-Shapelets, a novel framework that learns to predict gene expression dynamics by matching discriminative local temporal patterns. Instead of modeling entire irregular trajectories, G-Shapelets identify multiscale temporal motifs-or shapelets-that serve as robust and interpretable building blocks for prediction. This approach is naturally resilient to gaps and irregular time intervals because it operates on local subsequences. By learning a dictionary of these shapelets and measuring their distance to segments of an input gene's trajectory, our model constructs a feature representation that effectively captures dynamic behavior. Extensive experiments on the longitudinal viral immunization dataset (HR-VILAGE3K3M) demonstrate that G-Shapelets achieves$\mathbf{4. 8 \%}$lower mean squared error and statistically significant improvements over recurrent, convolutional, and transformer baselines. Furthermore, we show that the learned shapelets correspond to meaningful biological patterns, offering interpretability that is absent in blackbox models. This framework opens new avenues for accurate and interpretable modeling of dynamic transcriptomic systems.

Recent research has demonstrated the effectiveness of knowledge distillation in Domain Generalization. However, existing approaches often overlook domain-specific knowledge and rely on an offline distillation strategy, limiting the effectiveness of knowledge transfer. To address these limitations, we propose Balanced Online knowLedge Distillation (BOLD). BOLD leverages a multi-domain expert teacher model, with each expert specializing in a specific source domain, enabling the student to distill both domain-invariant and domain-specific knowledge. We incorporate the Pareto optimization principle and uncertainty weighting to balance these two types of knowledge, ensuring simultaneous optimization without compromising either. Additionally, BOLD employs an online knowledge distillation strategy, allowing the teacher and student to learn concurrently. This dynamic interaction enables the teacher to adapt based on student feedback, facilitating more effective knowledge transfer. Extensive experiments on seven benchmarks demonstrate that BOLD outperforms state-of-the-art methods. Furthermore, we provide theoretical insights that highlight the importance of domain-specific knowledge and the advantages of uncertainty weighting.

ABSTRACT Traditional frequent pattern mining algorithms often report an overwhelming number of patterns in large datasets, many of which are redundant. To address this issue, Minimum Description Length (MDL)‐based methods have been employed, which use data compression to capture a smaller yet significant set of patterns. However, finding a good set of patterns according to MDL involves a very large search space, and current MDL‐based techniques often suffer from long runtimes and find suboptimal solutions. To discover better sets of patterns in less time, this paper introduces GRIMP (a Genetic algoRIthm for coMpression‐based descriptive Pattern mining), a novel framework that combines a genetic algorithm with MDL‐based pattern selection. Multiple genetic algorithm variants are explored within the GRIMP framework, and their effectiveness is compared using a large number of datasets. Experimental results demonstrate that GRIMP consistently outperforms previous methods by achieving higher compression ratios, generating more representative itemsets, and requiring less time. Additionally, the extracted patterns improve downstream classification tasks, highlighting the ability of GRIMP to find more representative patterns within the data.

Repetitive Negative Sequential Patterns (RNSPs) can provide critical insights into the importance of sequences. However, most current RNSP mining methods require users to set an appropriate support threshold to obtain the expected number of patterns, which is a very difficult task for the users without prior experience . To address this issue, we propose a new algorithm, TK-RNSP, to mine the Top- K RNSPs with the highest support, without the need to set a support threshold. In detail, we achieve a significant breakthrough by proposing a series of definitions that enable RNSP mining to satisfy anti-monotonicity. Then, we propose a bitmap-based Depth-First Backtracking Search (DFBS) strategy to decrease the heavy computational burden by increasing the speed of support calculation. Finally, we propose the algorithm TK-RNSP in an one-stage process, which can effectively reduce the generation of unnecessary patterns and improve computational efficiency comparing to those two-stage process algorithms. To the best of our knowledge, TK-RNSP is the first algorithm to mine Top- K RNSPs. Extensive experiments on eight datasets show that TK-RNSP has better flexibility and efficiency to mine Top- K RNSPs.

Sequential pattern mining (SPM) with gap constraints (or repetitive SPM or tandem repeat discovery in bioinformatics) can find frequent repetitive subsequences satisfying gap constraints, which are called positive sequential patterns with gap constraints (PSPGs). However, classical SPM with gap constraints cannot find the frequent missing items in the PSPGs. To tackle this issue, this article explores negative sequential patterns with gap constraints (NSPGs). We propose an efficient NSPG-Miner algorithm that can mine both frequent PSPGs and NSPGs simultaneously. To effectively reduce candidate patterns, we propose a pattern join strategy with negative patterns which can generate both positive and negative candidate patterns at the same time. To calculate the support (frequency of occurrence) of a pattern in each sequence, we explore a NegPair algorithm that employs a key-value pair array structure to deal with the gap constraints and the negative items simultaneously and can avoid redundant rescanning of the original sequence, thus improving the efficiency of the algorithm. To report the performance of NSPG-Miner, 11 competitive algorithms and 11 datasets are employed. The experimental results not only validate the effectiveness of the strategies adopted by NSPG-Miner but also verify that NSPG-Miner can discover more valuable information than the state-of-the-art algorithms. Algorithms and datasets can be downloaded from https://github.com/wuc567/Pattern-Mining/tree/master/NSPG-Miner .

High utility itemset mining (HUIM) is one of the most compelling problems in data mining, extending frequent itemset mining (FIM) and serving as a crucial method for analyzing customer behavior. Many HUIM algorithms have been proposed to improve execution time and memory consumption. However, most assume that the profit is fixed for each item in a database, which is unrealistic. Some algorithms address products with unstable transaction profits but still need to run faster due to ineffective pruning strategies. Additionally, generalizing items into categories is often neglected. To address these issues, this paper considers a more practical database type that integrates unstable profits with a taxonomy of items. The proposed algorithm, CLHUN (Cross-level High Utility Itemset Mining in a Database with Unstable and Negative Profits), combines efficient techniques such as item sorting and tighter upper bounds to prune the search space. Furthermore, it introduces strategies to eliminate unpromising items during mining and reduce the number of transaction scans. Several experiments were conducted to evaluate the algorithm's performance. Results demonstrate that CLHUN is efficient with these techniques and strategies.

Deep learning methods often suffer performance degradation due to domain shift, where discrepancies exist between training and testing data distributions. Domain generalization mitigates this problem by leveraging information from multiple source domains to enhance model generalization capabilities for unseen domains. However, existing domain generalization methods typically present examples to the model in a random manner, overlooking the potential benefits of structured data presentation. To bridge this gap, we propose a novel learning strategy, Symmetric Self-Paced Learning (SSPL), for domain generalization. SSPL consists of a Symmetric Self-Paced training scheduler and a Gradient-based Difficulty Measure (GDM). Specifically, the proposed training scheduler initially focuses on easy examples, gradually shifting emphasis to harder examples as training progresses. GDM dynamically evaluates example difficulty through the gradient magnitude with respect to the example itself. Experiments across five popular benchmark datasets demonstrate the effectiveness of the proposed learning strategy.

Accurate analysis and classification of clinical datasets are crucial for understanding disease patterns, identifying risk factors and devising targeted interventions that ultimately contribute towards effective healthcare systems and improved patient outcomes. However, existing analysis and classification methods often fall short of effectively capturing complex sequential relationships within patient data and have limited interpretability. To overcome these challenges, we introduce SeqClin, a novel approach that utilizes frequent pattern mining to obtain valuable sequential information from clinical datasets. SeqClin first transforms clinical datasets into an appropriate format. Then, it employs sequential pattern mining algorithms to find frequent sequential patterns as well as rules of patient features in the datasets. These identified feature patterns and their respective values are then used for classification/detection. The performance of SeqClin is evaluated on four clinical datasets, where six classification models and evaluation metrics are employed for a comprehensive assessment. The obtained results show that the proposed approach surpassed previous approaches, with the extracted patterns and rules providing valuable insights into the key patient features and their values in clinical datasets.

The exponential growth of genomic data has posed significant challenges for lossless compression of genome sequences. While recent reference-free genome compressors have shown promising results, they often fail to fully leverage the inherent sequential structure of genome sequences, require substantial computational resources and lack (or have limited) interpretability. This paper presents a novel genome compression method that employs the Minimum Description Length (MDL) principle, which is based on the idea that the best model for a given dataset is the one that provides the shortest description of that dataset. The proposed compressor, called GMG (Genetic algorithm for MDL-based Genome compression), integrates a genetic algorithm to identify optimal k-mers (patterns) in a model to best compress the genome data. Experimental results across various datasets demonstrate that GMG outperforms state-of-the-art genome compressors in terms of bits-per-base compression and computational efficiency. Furthermore, it is demonstrated that the optimal patterns identified by GMG for compression can also be utilized for genome classification, offering a multifunctional advantage over previous compressors. GMG is freely available at github.com/MuhammadzohaibNawaz/GMG

Frequent itemset mining (FIM) is a widely used task that extracts frequently occurring itemsets from data. Plenty of deterministic algorithms are available for this daunting task. However, experimental studies have not considered that data heterogeneity significantly impacts the algorithms' performance, giving rise to unfair comparisons and biased conclusions. This paper seeks to advance by comparing cutting-edge algorithms using various frequency thresholds, considering the resulting data heterogeneity. An extensive experimental study is carried out, including the number of itemsets mined per second as the performance quality measure to compare algorithms. The experiments include defining eight metrics to quantify data heterogeneity, and their values vary the algorithms' performance. The results revealed that some techniques (hypercube decomposition and k-items machine) are essential to achieve excellent performance on any dataset, and most algorithms behave similarly well when they include those techniques. As a final important point, different threshold values produce dissimilar data subsets (data heterogeneity is not an immutable data characteristic), so a previous study on the database characteristics with a few minimum support thresholds could be beneficial to select the best-suited FIM algorithm beforehand.

The synthetic minority oversampling technique (SMOTE) is a popular algorithm to reduce the impact of class imbalance in building classifiers, and has received several enhancements over the past 20 years. SMOTE and its variants synthesize a number of minority-class sample points in the original sample space to alleviate the adverse effects of class imbalance. This approach works well in many cases, but problems arise when synthetic sample points are generated in overlapping areas between different classes, which further complicates classifier training. To address this issue, this paper proposes a novel generalization-oriented rather than imputation-oriented minority-class sample point generation algorithm, named overlapping minimization SMOTE (OM-SMOTE). This algorithm is designed specifically for binary imbalanced classification problems. OM-SMOTE first maps the original sample points into a new sample space by balancing sample encoding and classifier generalization. Then, OM-SMOTE employs a set of sophisticated minority-class sample point imputation rules to generate synthetic sample points that are as far as possible from overlapping areas between classes. Extensive experiments have been conducted on 32 imbalanced datasets to validate the effectiveness of OM-SMOTE. Results show that using OM-SMOTE to generate synthetic minority-class sample points leads to better classifier training performances for the naive Bayes, support vector machine, decision tree, and logistic regression classifiers than the 11 state-of-the-art SMOTE-based imputation algorithms. This demonstrates that OM-SMOTE is a viable approach for supporting the training of high-quality classifiers for imbalanced classification. The implementation of OM-SMOTE is shared publicly on the GitHub platform at https://github.com/luxuan123123/OM-SMOTE/ .

Sequential pattern mining (SPM) is an important branch of knowledge discovery that aims to mine frequent sub-sequences (patterns) in a sequential database. Various SPM methods have been investigated, and most of them are classical SPM methods, since these methods only consider whether or not a given pattern occurs within a sequence. Classical SPM can only find the common features of sequences, but it ignores the number of occurrences of the pattern in each sequence, i.e., the degree of interest of specific users. To solve this problem, this paper addresses the issue of repetitive nonoverlapping sequential pattern (RNP) mining and proposes the RNP-Miner algorithm. To reduce the number of candidate patterns, RNP-Miner adopts an itemset pattern join strategy. To improve the efficiency of support calculation, RNP-Miner utilizes the candidate support calculation algorithm based on the position dictionary. To validate the performance of RNP-Miner, 10 competitive algorithms and 20 sequence databases were selected. The experimental results verify that RNP-Miner outperforms the other algorithms, and using RNPs can achieve a better clustering performance than raw data and classical frequent patterns. All the algorithms were developed using the PyCharm environment and can be downloaded fromhttps://github.com/wuc567/Pattern-Mining/tree/master/RNP-Miner.

Recently, order-preserving pattern (OPP) mining, a new sequential pattern mining method, has been proposed to mine frequent relative orders in a time series. Although frequent relative orders can be used as features to classify a time series, the mined patterns do not reflect the differences between two classes of time series well. To effectively discover the differences between time series, this paper addresses the top-kcontrast OPP (COPP) mining and proposes a COPP-Miner algorithm to discover the top-kcontrast patterns as features for time series classification, avoiding the problem of improper parameter setting. COPP-Miner is composed of three parts: extreme point extraction to reduce the length of the original time series, forward mining, and reverse mining to discover COPPs. Forward mining contains three steps: group pattern fusion strategy to generate candidate patterns, the support rate calculation method to efficiently calculate the support of a pattern, and two pruning strategies to further prune candidate patterns. Reverse mining uses one pruning strategy to prune candidate patterns and consists of applying the same process as forward mining. Experimental results validate the efficiency of the proposed algorithm and show that top-kCOPPs can be used as features to obtain a better classification performance.

Image-text retrieval is one of the major tasks of cross-modal retrieval. Several approaches for this task map images and texts into a common space to create correspondences between the two modalities. However, due to the content (semantics) richness of an image, redundant secondary information in an image may cause false matches. To address this issue, this paper presents a semantic optimization approach, implemented as a Visual Semantic Loss (VSL), to assist the model in focusing on an image’s main content. This approach is inspired by how people typically annotate the content of an image by describing its main content. Thus, we leverage the annotated texts corresponding to an image to assist the model in capturing the main content of the image, reducing the negative impact of secondary content. Extensive experiments on two benchmark datasets (MSCOCO and Flickr30K) demonstrate the superior performance of our method. The code is available at: https://github.com/ZhangXu0963/VSL.

Node classification is the task of predicting the labels of unlabeled nodes in a graph. State-of-the-art methods based on graph neural networks achieve excellent performance when all labels are available during training. But in real-life, models are of ten applied on data with new classes, which can lead to massive misclassification and thus significantly degrade performance. Hence, developing open-set classification methods is crucial to determine if a given sample belongs to a known class. Existing methods for open-set node classification generally use transductive learning with part or all of the features of real unseen class nodes to help with open-set classification. In this paper, we propose a novel generative open-set node classification method, i.e., G2Pxy, which follows a stricter inductive learning setting where no information about unknown classes is available during training and validation. Two kinds of proxy unknown nodes, inter-class unknown proxies and external unknown proxies are generated via mixup to efficiently anticipate the distribution of novel classes. Using the generated proxies, a closed-set classifier can be transformed into an open-set one, by augmenting it with an extra proxy classifier. Under the constraints of both cross entropy loss and complement entropy loss, G2Pxy achieves superior effectiveness for unknown class detection and known class classification, which is validated by experiments on bench mark graph datasets. Moreover, G2Pxy does not have specific requirement on the GNN architecture and shows good generalizations.

High utility itemsets can reveal combinations of items that have a high profit, expense, or importance. Mining high utility itemsets in a database with$n$items generally results in a huge search space, composed of$2^{n}$itemsets, and heavy utility calculations for the explored itemsets. Previous algorithms using prefix tree structures perform two phases, namely candidate generation and testing. To avoid generating candidate itemsets, one-phase algorithms use list or hyper-link structures and have been proven to be superior to two-phase algorithms. However, it should be noted that a prefix tree is still an efficient structure for itemset mining problems, and especially algorithms using prefix trees such as FP-Growth have shown excellent performance for mining frequent itemsets. This paper proposes Hamm, a High-performance AlgorithM for Mining high utility itemsets. Hamm employs a novel TV (prefix Tree and utility Vector) structure and mines high utility itemsets in one phase without candidate generation. We also develop an efficient optimization which is incorporated into Hamm as a component. Using prefix trees and utility vectors, Hamm outperforms state-of-the-art algorithms on various databases in experiments. Experimental results also show that the proposed optimization remarkably reduces the search space and speeds up Hamm.

Discovering frequent trends in time series is a critical task in data mining. Recently, order-preserving matching was proposed to find all occurrences of a pattern in a time series, where the pattern is a relative order (regarded as a trend) and an occurrence is a sub-time series whose relative order coincides with the pattern. Inspired by the order-preserving matching, the existing order-preserving pattern (OPP) mining algorithm employs order-preserving matching to calculate the support, which leads to low efficiency. To address this deficiency, this paper proposes an algorithm called efficient frequent OPP miner (EFO-Miner) to find all frequent OPPs. EFO-Miner is composed of four parts: a pattern fusion strategy to generate candidate patterns, a matching process for the results of sub-patterns to calculate the support of super-patterns, a screening strategy to dynamically reduce the size of prefix and suffix arrays, and a pruning strategy to further dynamically prune candidate patterns. Moreover, this paper explores the order-preserving rule (OPR) mining and proposes an algorithm called OPR-Miner to discover strong rules from all frequent OPPs using EFO-Miner. Experimental results verify that OPR-Miner gives better performance than other competitive algorithms. More importantly, clustering and classification experiments further validate that OPR-Miner achieves good performance.

This paper proposes a novel and efficient method, called S-PDB, for the analysis and classification of Spike (S) protein structures of SARS-CoV-2 and other viruses/organisms in the Protein Data Bank (PDB). The method first finds and identifies protein structures in PDB that are similar to a protein structure of interest (SARS-CoV-2 S) via a protein structure comparison tool. The amino acid (AA) sequences of identified protein structures, downloaded from PDB, and their aligned amino acids (AAA) and secondary structure elements (ASSE), that are stored in three separate datasets, are then used for the reliable detection/classification of SARS-CoV-2 S protein structures. Three classifiers are used and their performance is compared by using six evaluation metrics. Obtained results show that two classifiers for text data (Multinomial Naive Bayes and Stochastic Gradient Descent) performed better and achieved high accuracy on the dataset that contains AAA of protein structures compared to the datasets for AA and ASSE, respectively.

Sequential rule mining (SRM) is an alternative to sequential pattern mining (SPM) when dealing with sequence data. SRM has a wide range of applications in numerous data analysis scenarios. Existing SRM algorithms usually discover the entire set of rules in the databases, which makes it not only difficult to analyze results because the discovered set is too large, but also does not consider the user’s expectations and background knowledge. To tackle this problem, researchers have explored related algorithms with different constraints according to their requirements. In this paper, we propose a flexible constraint-based SRM algorithm called ConSRM for discovering only the sequential rules within user-specified time bounds in a sequence database. This algorithm uses an efficient rule-growth method and develops corresponding constraints and pruning strategies to reduce the search space and speed up calculation. Comprehensive experiments were carried out on four real datasets to evaluate the performance (both effectiveness and efficiency) of ConSRM.

Graphs are a popular data type found in many domains. Numerous techniques have been proposed to find interesting patterns in graphs to help understand the data and support decision-making. However, there are generally two limitations that hinder their practical use: (1) they have multiple parameters that are hard to set but greatly influence results, (2) and they generally focus on identifying complex subgraphs while ignoring relationships between attributes of nodes. Graphs are a popular data type found in many domains. Numerous techniques have been proposed to find interesting patterns in graphs to help understand the data and support decision-making. However, there are generally two limitations that hinder their practical use: (1) they have multiple parameters that are hard to set but greatly influence results, (2) and they generally focus on identifying complex subgraphs while ignoring relationships between attributes of nodes. To address these problems, we propose a parameter-free algorithm named CSPM (Compressing Star Pattern Miner) which identifies star-shaped patterns that indicate strong correlations among attributes via the concept of conditional entropy and the minimum description length principle. Experiments performed on several benchmark datasets show that CSPM reveals insightful and interpretable patterns and is efficient in runtime. Moreover, quantitative evaluations on two real-world applications show that CSPM has broad applications as it successfully boosts the accuracy of graph attribute completion models by up to 30.68% and uncovers important patterns in telecommunication alarm data.

Structured sequences are a popular data representation, used to model complex data such as traffic networks. A key machine learning task for structured sequences is node classification, that is predicting the class labels of unlabeled nodes. Though many node classification models were proposed, they assume a closed world setting, that all class labels appear in the training data. But in the real-world, the presence of never-before-seen class labels in testing data can considerably degrade a classifier’s accuracy. A promising solution to this issue is to build classifiers for an open-world setting, where samples with unknown class labels are continuously observed such that training and testing data may have different class label spaces. Several approaches have been proposed for open-world learning problems in computer vision and natural language processing, but they cannot be applied directly to structured sequences due to the complexity of their non-Euclidean properties and their dynamic nature. This paper addresses this important research gap by proposing a novel Open-world Structured Sequence node Classification (OSSC) model, to learn from structured sequences in an open-world setting. OSSC captures the structural and temporal information via a GCN-based dynamic variational framework. A latent distribution sequence is learned for each node using both stochastic states and deterministic states, to capture the evolution of node attributes and topology, followed by a sampling process to generate node representations. An open-world classification loss is further adopted to ensure that node representations are sensitive to unknown classes. And a combination of Openmax and Softmax is utilized to recognize nodes from unknown classes and to classify others to one of the known classes. Experiments on real-world datasets show that the proposed OSSC method is capable of learning accurate open-world node classifiers from structured sequence data.

To alleviate the attribute independence assumption, this paper proposes an auto-encoding naive Bayesian classifier (AE-NBC) which is trained based on transformed independent condition-attributes using an auto-encoding strategy. First, an auto-encoding neural network (AENN) with single hidden-layer is created to map the original data set to a high-dimensional data space. To guarantee the independence among auto-encoding condition-attributes (AECAs), an efficient and convergent objective function is elaborated to optimize the input-layer and output-layer weights of the AENN by considering both the auto-encoding error and attribute independence degree. Second, independent component analysis is iteratively carried out to remove the redundant attributes in AECAs so that the optimal transformed condition-attributes (TCAs) are obtained for the training of the NBC. Experiments were conducted on 29 benchmark data sets to validate the rationality, feasibility, and effectiveness of AE- NBC. Experimental results show that the independence among TCAs is stronger than among the original condition-attributes, and that AE- NBC can obtain higher training and testing accuracies than six Bayesian classifiers.

Streaming data has become more common as our ability to collect data in real-time increases. A primary concern in dealing with data streams is concept drift, which describes changes in the underlying distribution of streaming data. Measuring drift severity is crucial for model adaptation. Drift severity can be a proxy in choosing concept drift adaptation strategies. Current methods measure drift severity by monitoring the changes in the learner performance or measuring the difference between data distributions. However, these methods cannot measure the drift severity if the ground truth labels are unavailable. Specifically, performance-based methods cannot measure marginal drift, and distribution-based methods cannot measure conditional drift. We propose a novel framework named Tree-based Drift Measurement (TDM) that measures both marginal and conditional drift without revisiting historical data. TDM measures the difference between tree classifiers by transforming them into sets of binary vectors. An experiment shows that TDM achieves similar performance to the state-of-the-art methods and provides the best trade-off between runtime and memory usage. A case study shows that the online learner performance can be improved by adapting different drift adaptation strategies based on the drift severity.

User guided proof development in interactive theorem proving is a manual and time consuming activity.For automating proof searching and optimization in a higher-order logic proof assistant, we provide two metaheuristic algorithms that are based on Fitness Dependent Optimizer (FDO) and Bat Algorithm (BA).In both metaheuristic algorithms, random proof sequences are first created from a population of frequently occurring proof steps that are discovered using pattern mining techniques.Created proof sequences are then evolved till their fitness matches the fitness of the original (or target) proof sequences.Experiments are performed to investigate the performance of the proposed algorithms on different HOL4 theories.Moreover, the proposed FDO and BA-based proof searching approaches are compared with Simulated Annealing (SA) and Genetic Algorithm (GA)based methods.Results show that BA performs best, followed by FDO and SA for proof finding and optimization in HOL4.

Significant place mining in spatiotemporal trajectory data is a key task for mobile pattern mining, useful for supporting location-aware services. State-of-the-art trajectory clustering algorithms utilize a density-based distance measure. However, some major problems with this approach are that (1) results are often inaccurate, especially on data of varying density, (2) the user must fine-tune many thresholds that are unintuitive to set, and (3) boundary points between clusters are often assigned to the wrong locations. Performance is also a major issue as many state-of-the-art algorithms have a very high time complexity. Motivated by these issues, this paper proposes an approach inspired by the data field theory and community detection. It is a graph-based significant place mining algorithm, called GB-SPM, for capturing and characterizing the essence of similarity between nodes. GB-SPM first applies a novel low index neighborhood velocity point filtration method to extract characteristic points. Then, a characteristic point index neighborhood is used to map them to graph nodes. In this way, the original problem is transformed into a community detection problem in complex community networks. Finally, a new edge weight metric is proposed to capture and characterize the nature of similarity between nodes. To evaluate clustering quality, we used the Silhouette (SI) for unannotated data to value inter-cluster separation and intra-cluster homogeneity. To evaluate mining effectiveness, we used Matthew’s correlation coefficient (MCC) for annotated data. Numerous experiments were carried out on real world datasets, and the accuracy and performance of the designed algorithm was compared with the state-of-the-art algorithms. Results show that GB-SPM improves on average SI by 13.9%, MCC by 20.7%, and runtime by 5.15 times.

This article presents a Bayesian attributebagging-based extreme learning machine (BAB-ELM)to handle high-dimensional classification and regression problems. First, thedecision-making degree (DMD)of a condition attribute is calculated based on the Bayesian decision theory, i.e., the conditional probability of the condition attribute given the decision attribute. Second, the condition attribute with the highest DMD is put into thecondition attribute group (CAG)corresponding to the specific decision attribute. Third, thebagging attribute groups (BAGs)are used to train an ensemble learning model ofextreme learning machines (ELMs).Each base ELM is trained on a BAG which is composed of condition attributes that are randomly selected from the CAGs. Fourth, the information amount ratios of bagging condition attributes to all condition attributes is used as the weights to fuse the predictions of base ELMs in BAB-ELM. Exhaustive experiments have been conducted to compare the feasibility and effectiveness of BAB-ELM with seven other ELM models, i.e., ELM, ensemble-based ELM (EN-ELM), voting-based ELM (V-ELM), ensemble ELM (E-ELM), ensemble ELM based on multi-activation functions (MAF-EELM), bagging ELM, and simple ensemble ELM. Experimental results show that BAB-ELM is convergent with the increase of base ELMs and also can yield higher classification accuracy and lower regression error for high-dimensional classification and regression problems.

Nonoverlapping sequential pattern mining is an important type of sequential pattern mining (SPM) with gap constraints, which not only can reveal interesting patterns to users but also can effectively reduce the search space using the Apriori (anti-monotonicity) property. However, the existing algorithms do not focus on attributes of interest to users, meaning that existing methods may discover many frequent patterns that are redundant. To solve this problem, this article proposes a task called nonoverlapping three-way sequential pattern (NTP) mining, where attributes are categorized according to three levels of interest: strong, medium, and weak interest. NTP mining can effectively avoid mining redundant patterns since the NTPs are composed of strong and medium interest items. Moreover, NTPs can avoid serious deviations (the occurrence is significantly different from its pattern) since gap constraints cannot match with strong interest patterns. To mine NTPs, an effective algorithm is put forward, called NTP-Miner, which applies two main steps: support (frequency occurrence) calculation and candidate pattern generation. To calculate the support of an NTP, depth-first and backtracking strategies are adopted, which do not require creating a whole Nettree structure, meaning that many redundant nodes and parent–child relationships do not need to be created. Hence, time and space efficiency is improved. To generate candidate patterns while reducing their number, NTP-Miner employs a pattern join strategy and only mines patterns of strong and medium interest. Experimental results on stock market and protein datasets show that NTP-Miner not only is more efficient than other competitive approaches but can also help users find more valuable patterns. More importantly, NTP mining has achieved better performance than other competitive methods in clustering tasks. Algorithms and data are available at: https://github.com/wuc567/Pattern-Mining/tree/master/NTP-Miner .

High utility itemset mining is an important model in data mining. It involves discovering all itemsets in a quantitative transactional database that satisfy a user-specified minimum utility (minUtil) constraint. MinUtil controls the minimum value that an itemset must maintain in a database. Since the model evaluates an itemset’s interestingness using only the minUtil constraint, it implicitly assumes that all items in the database have similar utility values. However, some items have high utility, while others may have relatively low utility in a database. If minUtil is set too high, the user will miss all itemsets containing low utility items. To find itemsets that involve both high and low utility items, minUtil has to be set very low. However, this may cause a combinatorial explosion as the items with high utility may combine with others in all possible ways. This dilemma is called the low utility item problem. This paper proposes a flexible model of relative high utility itemset to address this problem. We introduce a new null-invariant measure, called utility ratio, to evaluate the interestingness of an itemset in the database. We also present a fast single scan algorithm to find all desired itemsets in the database. Experimental results demonstrate that the proposed algorithm is efficient. Finally, a case study on Yahoo! JAPAN retail data shows that the proposed model is useful.

High Utility Itemset Mining (HUIM) aims at finding all sets of items that have high importance in a database, as measured by a utility function. Although HUIM has many applications, a key limitation is that the discovered patterns often have an unstable utility over time. For example, while a set of products may yield a high utility (profit) over a year, that utility may fluctuate from weeks to weeks. To discover patterns that have a stable utility and hence that are more suitable for decision-making, this paper redefines HUIM as the task of discovering Stable High Utility Itemsets (StableHUI). An efficient tree-based and pattern-growth algorithm named Stable-Growth is proposed to extract all the StableHUI. Several experiments on two real-world datasets and two synthetic datasets show that Stable-Growth is up to 60% faster than a baseline and that it can filter out numerous unstable HUI.

This paper proposes a novel method for video quality evaluation based on machine learning technique. The current research deals with the correct interpretation of objective video quality evaluation (Quality of Service – QoS) in relation to subjective end-user perception (Quality of Experience – QoE), typically expressed by mean opinion score (MOS). Our method allows us to interconnect results obtained from video objective and subjective assessment methods in the form of a neural network (computing model inspired by biological neural networks). So far, no unified interpretation scale has been standardized for both approaches, therefore it is difficult to determine the level of end-user satisfaction obtained from the objective assessment. Thus, contribution of the proposed method lies in description of the way to create a hybrid metric that delivers fast and reliable subjective score of perceived video quality for internet television (IPTV) broadcasting companies.

In industrial areas, most of databases are dynamic databases, and the volume of the databases has grown with the passage of time. Especially, pattern mining for incremental database needs different approaches from static database because the profit or the accuracy of the previously inserted data can be reduced. Since data is time- sensitive, the recent data has a relatively higher value than the old data. In this paper, we suggest the damped window based average utility driven data analytics for intelligent systems, which the damped window reflects the importance according to the arrival time of the transactions. The proposed mining approach adopts novel data structure, which modify the importance of item as the passage of time, and it improves mining efficiency with several pruning strategies and without generating candidate patterns. To evaluate the performance of the proposed mining approach, we conducted various experiments using several real and synthetic data sets. The result of the experiments presented that the suggested method performs better in terms of runtime and memory usage than the other state-of-the-art mining techniques. Moreover, through the scalability experiments, which changed the number of different items or transactions, we verified that the proposed algorithm maintained a stable performance under various environmental changes.

High-utility sequential pattern (HUSP) mining is an emerging topic in the field of knowledge discovery in databases. It consists of discovering subsequences that have a high utility (importance) in sequences, which can be referred to as HUSPs. HUSPs can be applied to many real-life applications, such as market basket analysis, e-commerce recommendations, click-stream analysis, and route planning. Several algorithms have been proposed to efficiently mine utility-based useful sequential patterns. However, due to the combinatorial explosion of the search space for low utility threshold and large-scale data, the performances of these algorithms are unsatisfactory in terms of runtime and memory usage. Hence, this article proposes an efficient algorithm for the task of HUSP mining, called HUSP mining with UL-list (HUSP-ULL). It utilizes a lexicographic q -sequence (LQS)-tree and a utility-linked (UL)-list structure to quickly discover HUSPs. Furthermore, two pruning strategies are introduced in HUSP-ULL to obtain tight upper bounds on the utility of the candidate sequences and reduce the search space by pruning unpromising candidates early. Substantial experiments on both real-life and synthetic datasets showed that HUSP-ULL can effectively and efficiently discover the complete set of HUSPs and that it outperforms the state-of-the-art algorithms.

Knowledge extraction from database is the fundamental task in database and data mining community, which has been applied to a wide range of real-world applications and situations. Different from the support-based mining models, the utility-oriented mining framework integrates the utility theory to provide more informative and useful patterns. Time-dependent sequence data are commonly seen in real life. Sequence data have been widely utilized in many applications, such as analyzing sequential user behavior on the Web, influence maximization, route planning, and targeted marketing. Unfortunately, all the existing algorithms lose sight of the fact that the processed data not only contain rich features (e.g., occur quantity, risk, and profit), but also may be associated with multi-dimensional auxiliary information, e.g., transaction sequence can be associated with purchaser profile information. In this article, we first formulate the problem of utility mining across multi-dimensional sequences, and propose a novel framework named MDUS to extract Multi-Dimensional Utility-oriented Sequential useful patterns. To the best of our knowledge, this is the first study that incorporates the time-dependent sequence-order, quantitative information, utility factor, and auxiliary dimension. Two algorithms respectively named MDUS EM and MDUS SD are presented to address the formulated problem. The former algorithm is based on database transformation, and the later one performs pattern joins and a searching method to identify desired patterns across multi-dimensional sequences. Extensive experiments are carried on six real-life datasets and one synthetic dataset to show that the proposed algorithms can effectively and efficiently discover the useful knowledge from multi-dimensional sequential databases. Moreover, the MDUS framework can provide better insight, and it is more adaptable to real-life situations than the current existing models.

The main purpose of data mining and analytics is to find novel, potentially useful patterns that can be utilized in real-world applications to derive beneficial knowledge. For identifying and evaluating the usefulness of different kinds of patterns, many techniques and constraints have been proposed, such as support, confidence, sequence order, and utility parameters (e.g., weight, price, profit, quantity, satisfaction, etc.). In recent years, there has been an increasing demand for utility-oriented pattern mining (UPM, or called utility mining). UPM is a vital task, with numerous high-impact applications, including cross-marketing, e-commerce, finance, medical, and biomedical applications. This survey aims to provide a general, comprehensive, and structured overview of the state-of-the-art methods of UPM. First, we introduce an in-depth understanding of UPM, including concepts, examples, and comparisons with related concepts. A taxonomy of the most common and state-of-the-art approaches for mining different kinds of high-utility patterns is presented in detail, including Apriori-based, tree-based, projection-based, vertical-/horizontal-data-format-based, and other hybrid approaches. A comprehensive review of advanced topics of existing high-utility pattern mining techniques is offered, with a discussion of their pros and cons. Finally, we present several well-known open-source software packages for UPM. We conclude our survey with a discussion on open and practical challenges in this field.

Consumer behavior plays a very important role in economics and targeted marketing. However, understanding economic consumer behavior is quite challenging, such as finding credible and reliable information on product profitability. Different from frequent pattern mining, utility-oriented mining integrates utility theory and data mining. Utility mining is a useful tool for understanding economic consumer behavior. Traditional algorithms for mining high-utility patterns (HUPs) applies a single/uniform minimum utility threshold ( minutil ) to obtain the set of HUPs, but in some real-life circumstances, some specific products may bring lower utilities compared with others, but their profit may offer some vital information. If minutil is set high, the patterns with low minutil are missed; if minutil is set low, the number of patterns becomes unmanageable. In this article, an efficient one-phase utility-oriented pattern mining algorithm, called HIMU, is proposed for mining HUPs with varied item-specific minimum utility. A novel tree structure called a multiple item utility set-enumeration tree (MIU-tree) and the global sorted and the conditional downward closure properties are introduced in HIMU. In addition, we extended the compact utility-list structure to keep the necessary information, and thus this one-phase HIMU model greatly reduces the computational costs and memory requirements. Moreover, two pruning strategies are then extended to enhance the performance. We conducted extensive experiments in several synthetic and real-world datasets; the results indicate that the designed one-phase HIMU algorithm can address the “ rare item problem ” and has better performance than the state-of-the-art algorithms in terms of runtime, memory usage, and scalability. Furthermore, the enhanced algorithms outperform the non-optimized HIMU approach.

Frequent pattern mining is an important model in data mining. It involves finding all patterns in a transactional database that satisfy the user-specified minimum support (minSup) constraint. The minSup controls the minimum number of transactions that a pattern must cover in a transactional database. Since only minSup is used to evaluate a pattern's interestingness, the frequent pattern model implicitly assumes that spatial information of the items will not impact the interestingness of a pattern in the database. This assumption limits the applicability of the frequent pattern model in many real-world applications. It is because patterns whose items are close to each other are typically more attractive to the user than the patterns whose items are far from each other in a coordinate system. With this motivation, this paper proposes a novel model of frequent spatial pattern that may exist in a spatiotemporal database. An efficient pattern-growth algorithm, called Frequent Spatial Pattern-growth (FSP-growth), has also been presented to mine all desired patterns in a database. Experimental results demonstrate that our algorithm is efficient. The usefulness of the proposed patterns has also been shown with a real-world application.

Mining useful patterns from varied types of databases is an important research topic, which has many real-life applications. Most studies have considered the frequency as sole interestingness measure to identify high-quality patterns. However, each object is different in nature. The relative importance of objects is not equal, in terms of criteria, such as the utility, risk, or interest. Besides, another limitation of frequent patterns is that they generally have a low occupancy, that is, they often represent small sets of items in transactions containing many items and, thus, may not be truly representative of these transactions. To extract high-quality patterns in real-life applications, this paper extends the occupancy measure to also assess the utility of patterns in transaction databases. We propose an efficient algorithm named high-utility occupancy pattern mining (HUOPM). It considers user preferences in terms of frequency, utility, and occupancy. A novel frequency-utility tree and two compact data structures, called the utility-occupancy list and frequency-utility table, are designed to provide global and partial downward closure properties for pruning the search space. The proposed method can efficiently discover the complete set of high-quality patterns without candidate generation. Extensive experiments have been conducted on several datasets to evaluate the effectiveness and efficiency of the proposed algorithm. Results show that the derived patterns are intelligible, reasonable, and acceptable, and that HUOPM with its pruning strategies outperforms the state-of-the-art algorithm, in terms of runtime and search space, respectively.

High-Utility Sequential Pattern Mining (HUSPM) has become an emerging issue in recent decades since it reveals more information such as the utility and sequence factors for knowledge discovery. For the previous works, many algorithms were presented to speed up the mining performance regarding a single machine with small datasets. In real-world applications, the size of dataset can be collected from many places or devices, such as PC, Internet of Things (IoT), mobile devices, and shopping malls, among others. It is necessary to build an efficient model to handle the big dataset for HUSPM. In this paper, we present a four-stages MapReduce framework based on the Spark platform for mining the high-utility sequential patterns from a very large database. From the experimental results, we then can observe that the designed model outperforms the state-of-the-art approaches for handling the very big dataset.

Comparing to frequent itemset mining (FIM), utility-pattern mining receives increasing attention in the field of data mining recently. With the flourishing development of utility-pattern mining, most studies focused on the efficiency problem by considering the efficient data structure to compress the original data and pruning strategies to reduce the search space for knowledge discovery. However, those approaches can only handle the binary situation, thus the discovered knowledge cannot be represented as the linguistic variables. Previous works have addressed this problem by introducing the generic approaches to find the high fuzzy utility itemsets in a small database. In real-world situations, the dataset may be very large, and it is costly to mine all the required information from a very large database. In this paper, we first present a HFUI-GA framework to discover the high fuzzy utility itemsets in a limited time. Several improvement strategies are also proposed to speed up the evolutionary progress. Experiments are then conducted to show the performance of the variants of the designed HFUI-GA framework in terms of number of the discovered high fuzzy utility itemsets (HFUIs) and the results are convincing to show that the designed GA-based HFUI-GA framework is a promising solution to mine for HFUIs.

Frequent Itemset Mining (FIM) is an important but very time-consuming data mining task. As a result, traditional FIM algorithms are often not scalable to large databases. To address this issue, several metaheuristics have been developed in recent years to find good approximate solutions to the FIM problem. It was shown that such approaches can be much more efficient than exact algorithms. However, metaheuristics often have long runtimes on massive datasets and the quality of their solutions can be improved. To address this issue, this paper proposes a parallel framework called CFIM (Cluster for Frequent Itemset Mining) for metaheuristic-based FIM. It accelerates FIM by using multiple cluster workers. The proposed approach partitions a transactional database and the set of all items at the level of cluster workers. The itemset generation process is performed by each worker, which then send results to a master node. This latter performs a merging step to only keep high quality itemsets by considering their frequency and diversification. Three metaheuristics (GA, PSO and BSO) are integrated in this framework to yield three novel metaheuristics (CGA, CPSO and CBSO). Extensive experiments show that CPSO outperforms CGA, CBSO, and state-of-the-art high performance computing FIM approaches.

In recent decades, data protection (PPDM), which not only hides information, but also provides information that is useful to make decisions, has become a critical concern. We present a sanitization algorithm with the consideration of four side effects based on multi-objective PSO and hierarchical clustering methods to find optimized solutions for PPDM. Experiments showed that compared to existing approaches, the designed sanitization algorithm based on the hierarchical clustering method achieves satisfactory performance in terms of hiding failure, missing cost, and artificial cost.

In this paper, we address the limitation for mining of high-utility sequential-pattern mining from uncertain databases and present a probabilistic high average-utility sequential pattern mining framework for discovering the set of probabilistic high average-utility sequential patterns from uncertain databases. A level-wise algorithm and three pruning strategies are introduced to mine the set of the desired patterns. Several experiments are then evaluated to show that the proposed algorithm achieves promising performance.

In real-world databases, high utility itemset (HUI) is an important class of regularities. Most previous studies have focused on mining HUIs in transactional databases and did not consider the spatiotemporal characteristics of items. In this study, a more flexible model of spatial HUIs (SHUIs) that exist in spatiotemporal databases is proposed. In a spatiotemporal database (STD), an itemset is said to be an SHUI if its utility is not less than a user-specified minimum utility and the distance between any two of its items is not more than a user-specified maximum distance. Identifying SHUIs is very challenging because the generated itemsets do not satisfy the anti-monotonic property. In this study, we present two novel pruning techniques for reducing computational costs. Moreover, a fast single scan algorithm is presented for effectively evaluating all SHUIs in a STD. Furthermore, two case studies are presented, in which the proposed model is used to identify useful information in traffic congestion data and air pollution data.

Association Rule Mining (ARM) is a fundamental data mining task that is time-consuming on big datasets. Thus, developing new scalable algorithms for this problem is desirable. Recently, Bee Swarm Optimization (BSO)-based meta-heuristics were shown effective to reduce the time required for ARM. But these approaches were applied only on small or medium scale databases. To perform ARM on big databases, a promising approach is to design parallel algorithms using the massively parallel threads of a GPU processor. While some GPU-based ARM algorithms have been developed, they only benefit from GPU parallelism during the evaluation step of solutions obtained by the BSO-metaheuristics. This paper improves this approach by parallelizing the other steps of the BSO process (diversification and intensification). Based on these novel ideas, three novel algorithms are presented, i) DRGPU (Determination of Regions on GPU), ii) SAGPU (Search Area on GPU, and, iii) ALLGPU (All steps on GPU). These solutions are analyzed and empirically compared on benchmark datasets. Experimental results show that ALLGPU outperforms the three other approaches in terms of speed up. Moreover, results confirm that ALLGPU outperforms the state-of-the-art GPU-based ARM approaches on big ARM databases such as the Webdocs dataset. Furthermore, ALLGPU is extended to mine big frequent graphs and results demonstrate its superiority over the state-of-the-art D-Mine algorithm for frequent graph mining on the large Pokec social network dataset.

The natural language processing (NLP) denotes a technique used to process data such as text and speech. Some of the fundamental research in NLP includes the named entity recognition, which recognizes the named entities (i.e., persons and companies) from texts, the semantic parsing, which converts a natural language utterance to a logical form, and the co-reference resolution, which extracts the nouns (including pronouns and noun phrases) pointing to the same reference body. In this paper, we focus on the mention extraction and classification, proposing a neural-encoded mention-hypergraph model named the BILU-NEMH to extract the mention entities from a content. The proposed BILU-NEMH model combines a mention hypergraph model with the encoding schema and neural network. The proposed model can effectively capture the overlapping mention entities of an unbounded length. The proposed model was verified by the experiments, and the obtained experimental results showed that the proposed model achieved better performance and greater effectiveness than the existing related models on most standard datasets.

With the growing popularity of shared resources, large volumes of complex data of different types are collected automatically. Traditional data mining algorithms generally have problems and challenges including huge memory cost, low processing speed, and inadequate hard disk space. As a fundamental task of data mining, sequential pattern mining (SPM) is used in a wide variety of real-life applications. However, it is more complex and challenging than other pattern mining tasks, i.e., frequent itemset mining and association rule mining, and also suffers from the above challenges when handling the large-scale data. To solve these problems, mining sequential patterns in a parallel or distributed computing environment has emerged as an important issue with many applications. In this article, an in-depth survey of the current status of parallel SPM (PSPM) is investigated and provided, including detailed categorization of traditional serial SPM approaches, and state-of-the art PSPM. We review the related work of PSPM in details including partition-based algorithms for PSPM, apriori-based PSPM, pattern-growth-based PSPM, and hybrid algorithms for PSPM, and provide deep description (i.e., characteristics, advantages, disadvantages, and summarization) of these parallel approaches of PSPM. Some advanced topics for PSPM, including parallel quantitative/weighted/utility SPM, PSPM from uncertain data and stream data, hardware acceleration for PSPM, are further reviewed in details. Besides, we review and provide some well-known open-source software of PSPM. Finally, we summarize some challenges and opportunities of PSPM in the big data era.

Mining High Average-Utility Itemsets (HAUIs) in a quantitative database is an extension of the traditional problem of frequent itemset mining, having several practical applications. Discovering HAUIs is more challenging than mining frequent itemsets using the traditional support model since the average-utilities of itemsets do not satisfy the downward-closure property. To design algorithms for mining HAUIs that reduce the search space of itemsets, prior studies have proposed various upper-bounds on the average-utilities of itemsets. However, these algorithms can generate a huge amount of unpromising HAUI candidates, which result in high memory consumption and long runtimes. To address this problem, this paper proposes four tight average-utility upper-bounds, based on a vertical database representation, and three efficient pruning strategies. Furthermore, a novel generic framework for comparing average-utility upper-bounds is presented. Based on these theoretical results, an efficient algorithm named dHAUIM is introduced for mining the complete set of HAUIs. dHAUIM represents the search space and quickly compute upper-bounds using a novel IDUL structure. Extensive experiments show that dHAUIM outperforms four state-of-the-art algorithms for mining HAUIs in terms of runtime on both real-life and synthetic databases. Moreover, results show that the proposed pruning strategies dramatically reduce the number of candidate HAUIs.

Successive location recommendation has recently emerged as an important service in Location-Based Social Networks (LBSNs). It aims at recommending the next location(s) to visit to a user given its current and previous locations. Although several recommenders have been proposed, only few works have considered the sequential correlations among locations in addition to other influential factors in recommendation. In this paper, a novel sequential rule mining-based approach called Location Recommender (LocRec) is proposed. Our proposal is designed to perform successive location recommendation for users of LBSN by considering sequential, social and temporal influence factors. The proposed approach first extracts the set of location sequences from mobility data and then generates recommendation rules from it. Based on the concept of recommendation influence factor, two rule-based recommenders are designed, namely temporal-based and social-based recommenders. An experimental evaluation was conducted on a real large-scale LBSN dataset to compare the performance of the proposed recommenders. Obtained results show that the tolerance to order-variations and the use of a window constraint enhance the performance of LocRec. They also depict that LocRec outperforms classical sequential-based models for successive location recommendation.

Recent studies have shown that radio data transmission is the most power consuming operation in power-constrained wireless sensor networks (WSNs). The power needed to transmit a single bit can be equal to that required for processing more than one thousand bits. Hence, increasing data processing to reduce data radio transmissions can considerably decrease power consumption. Several approaches have been designed to reduce data transmissions. This paper presents a novel approach, which combines partitioning clustering with the ARIMA algorithm. A performance comparison of the proposed approach with other data transmission algorithms, such as prediction using simple autoregressive and Adaptive Distributed Data Gathering (ADiDaG), is presented. Data from a real WSN, composed of more than 20,000 water meters in the City of Moncton, is used as a case study. Experimental results showed that the designed approach outperforms compared methods in terms of power reduction.

Discovering workflow patterns in event-logs is important for many organizations to understand and optimize organizational processes. Although numerous algorithms have been proposed in the literature to discover patterns in sequences of symbols, most of them are inadequate to discover patterns in rich event-log data. In this paper, motivated by the analysis of patient pathways in the health domain, a rich type of event logs, called activity-cost event logs, is considered where each event is associated with a cost. The paper formalizes the problem of mining interesting low-cost patterns in these logs by combining novel concepts of penalties (activity costs) and consistency of patterns, with traditional measures of confidence, length, and time. Furthermore, to extract these patterns efficiently from event logs, an algorithm named TWINCLE (Time-WINdow, Cost and LEngth constrained sequential rule mining) is proposed. Experiments carried out on benchmark datasets and real-life healthcare event logs show that proposed algorithm is efficient and can discover interesting patterns.

Detecting students' attention in class provides key information to teachers to capture and retain students' attention. Traditionally, such information is collected manually by human observers. Wearable devices, which have received a lot of attention recently, are rarely discussed in this field. In view of this, we propose a multimodal system which integrates a head-motion module, a pen-motion module, and a visual-focus module to accurately analyze students' attention levels in class. These modules collect information via cameras, accelerometers, and gyroscopes integrated in wearable devices to recognize students' behaviors. From these behaviors, attention levels are inferred for various time periods using a rule-based approach and a data-driven approach. The former infers a student's attention states using user-defined rules, while the latter relies on hidden relationships in the data. Extensive experimental results show that the proposed system has excellent performance and high accuracy. To the best of our knowledge, this is the first study on attention level inference in class using wearable devices. The outcome of this research has the potential of greatly increasing teaching and learning efficiency in class.

Frequent itemset mining (FIM) is one of the most common data mining techniques, which is based on the analysis of the occurrence frequencies of items in transactions. However, it is inapplicable in real-life situations since customers may purchase several units of the same item and all items may no t have the same unit profits. High-utility itemset mining (HUIM) was designed to consider both the quantities and unit profits of items in databases, and has become an emerging and critical research topic in recent decades. The SKYMINE approach was proposed to mine the skyline frequent-utility patterns (SFUPs), by considering both the utility and the occurrence frequencies of items. A SFUP is a non-dominated itemset, where the dominance relationship between itemsets is based on the utility and frequency measures. Mining SFUPs using the SKYMINE algorithm and its (UP)-tree structure requires, however, long execution times. In this paper, we propose a more efficient algorithm named skyline frequency-utility (SFU)-Miner to mine the SFUPs, utilizing the utility-list structure. This latter structure is used to efficiently calculate the actual utilities of itemsets without generating candidates, contrarily to the SKYMINE algorithm and its UP-tree structure. Besides, an array called utility-max (umax) is further developed to keep information about the maximal utility for each occurrence frequency, which can be used to greatly reduce the amount of itemsets considered for directly mining the SFUPs. This property can be used to efficiently find the non-dominated itemsets based on the utility and frequency measures. Substantial experiments have been carried out to evaluate the proposed algorithm’s performance. Results have shown that SFU-Miner outperforms the state-of-the-art SKYMINE algorithm for SFUP mining in terms of runtime, memory consumption, number of candidates, and scalability.

Privacy preserving data mining (PPDM) is an emerging research problem that has become critical in the last decades. PPDM consists of hiding sensitive information to ensure that it cannot be discovered by data mining algorithms. Several PPDM algorithms have been developed. Most of them are designed for hiding sensitive frequent itemsets or association rules. Hiding sensitive information in a database can have several side effects such as hiding other non-sensitive information and introducing redundant information. Finding the set of itemsets or transactions to be sanitised that minimises side effects is an NP-hard problem. In this paper, a genetic algorithm (GA) using transaction deletion is designed to hide sensitive high-utility itemsets for PPUM. A flexible fitness function with three adjustable weights is used to evaluate the goodness of each chromosome for hiding sensitive high-utility itemsets. To speed up the evolution process, the pre-large concept is adopted in the designed algorithm. It reduces the number of database scans required for verifying the goodness of an evaluated chromosome. Substantial experiments are conducted to compare the performance of the designed GA approach (with/without the pre-large concept), with a GA-based approach relying on transaction insertion and a non-evolutionary algorithm, in terms of execution time, side effects, database integrity and utility integrity. Results demonstrate that the proposed algorithm hides sensitive high-utility itemsets with fewer side effects than previous studies, while preserving high database and utility integrity.

In the past, several algorithms were developed to mine fuzzy frequent itemsets (FFIs) in which each item is represented at most one linguistic term based on maximum scalar cardinality. In real-life situations, multiple fuzzy linguistic terms instead of the single one can, however, produce more useful and meaningful fuzzy association rules. The Apriori-based algorithm was developed to mine multiple fuzzy frequent itemsets (MFFIs), which requires to generate the amounts of candidates and determine them in a level-wise way. In this paper, a fuzzy-list-based (FL)-Miner algorithm is developed to mine the complete set of MFFIs without candidate generation. Two efficient pruning strategies are also developed to reduce the search space, thus speeding up the mining process to directly discover the MFFIs. Experiments are conducted to show the performance of the proposed approaches compared to the state-of-the-art level-wise algorithm in terms of execution time and memory usage.

Mining of high-utility itemsets in transactional databases is emerging topic in recent years since it can be used to reveal more information for decision making, which has been widely used in many real-life applications. For the traditional high-utility itemset mining (HUIM), only the utility values of the itemsets are considered without timestamps or periodic constraints. In this paper, we present a new short periodic high-utility itemset mining (SPHUIM) to mine the set of short periodic high-utility itemsets (SPHUIs) by considering both the period and utility measures. A baseline two-phase SPHUI-TP algorithm is first presented to mine SPHUIs in level-wise manner. To reduce the search space of SPHUI-TP algorithm, two pruning strategies are also developed to speed up the mining performance of the SPHUI-TP algorithm. Substantial experiments both on real-life and synthetic datasets showed the efficiency and effectiveness of the designed approaches.

Mining sequential patterns (SPs) is a popular data mining task, which consists in finding interesting, unexpected, and useful patterns in sequence databases. It has several applications in many domains. However, most sequential pattern mining algorithms assume that databases are static, i.e. that they do not change over time. But in real-word applications, sequences are often modified. Thus, it is an important issue to design algorithms for updating SPs in a dynamic database environment. Although some algorithms have been proposed to maintain SPs in dynamic databases, these algorithms may have poor performance, especially when databases contain long sequences or a large number of sequences. This paper addresses this issue by proposing a novel dynamic mining approach named PreFUSP-TREE-MOD to address the problem of maintaining and updating discovered SPs when sequences in a database are modified. The proposed approach adopts the previously proposed pre-large concept using two support thresholds, to avoid scanning the database when possible, for updating the set of discovered patterns. Due to the pruning properties of the pre-large concept, the PreFUSP-TREE-MOD maintenance algorithm can effectively reduce the cost of database scans to maintain and update the built FUSP-tree for sequence modification. When the number of modified sequences is less than the safety bound of the pre-large concept, the proposed maintenance algorithm outperforms traditional SPM algorithms in batch mode, and the state-of-the-art maintenance algorithm in terms of execution time and number of tree nodes.

High utility itemsets (HUIs) mining is an emerging topic in data mining, which refers to discovering all itemsets having a utility meeting a user-specified minimum utility threshold min_util. However, setting min_util appropriately is a difficult problem for users. Generally speaking, finding an appropriate minimum utility threshold by trial and error is a tedious process for users. If min_util is set too low, too many HUIs will be generated, which may cause the mining process to be very inefficient. On the other hand, if min_util is set too high, it is likely that no HUIs will be found. In this paper, we address the above issues by proposing a new framework for top-k high utility itemset mining, where k is the desired number of HUIs to be mined. Two types of efficient algorithms named TKU (mining Top-K Utility itemsets) and TKO (mining Top-K utility itemsets in One phase) are proposed for mining such itemsets without the need to set min_util. We provide a structural comparison of the two algorithms with discussions on their advantages and limitations. Empirical evaluations on both real and synthetic datasets show that the performance of the proposed algorithms is close to that of the optimal case of state-of-the-art utility mining algorithms.

In recent years, mining high-utility itemsets (HUIs) has become as a key topic in data mining. However, most of the developed algorithms assume the unrealistic situations that unit profits of items remain unchanged over time. But in real-life situations, the profit of an item or itemset varies as a function of cost prices, sales prices and sales strategies. In this paper, a novel framework for mining HUIs with two algorithms under various Discount strategies (HUID) are introduced. HUID-tp is based on various discount strategies and a novel downward closure property to mine the complete set of HUIs. HUID-Miner is an algorithm relying on a compact data structure (Positive-and-Negative Utility-list, PNU-list) and new pruning strategies to efficiently discover HUIs without candidate generation, while considerably reducing the size of the search space. Furthermore, a strategy named Estimated Utility Co-occurrence Strategy which stores the relationships between 2-itemsets is also adopted in the proposed improvement HUID-EMiner algorithm to speed up computation. An extensive experimental study carried on several real-life datasets shows the performance of the proposed algorithms.

Sequential rule mining is an important data mining problem with multiple applications. An important limitation of algorithms for mining sequential rules common to multiple sequences is that rules are very specific and therefore many similar rules may represent the same situation. This can cause three major problems: (1) similar rules can be rated quite differently, (2) rules may not be found because they are individually considered uninteresting, and (3) rules that are too specific are less likely to be used for making predictions. To address these issues, we explore the idea of mining “partially-ordered sequential rules” (POSR), a more general form of sequential rules such that items in the antecedent and the consequent of each rule are unordered. To mine POSR, we propose the RuleGrowth algorithm, which is efficient and easily extendable. In particular, we present an extension (TRuleGrowth) that accepts a sliding-window constraint to find rules occurring within a maximum amount of time. A performance study with four real-life datasets show that RuleGrowth and TRuleGrowth have excellent performance and scalability compared to baseline algorithms and that the number of rules discovered can be several orders of magnitude smaller when the sliding-window constraint is applied. Furthermore, we also report results from a real application showing that POSR can provide a much higher prediction accuracy than regular sequential rules for sequence prediction.

Mining high utility itemsets (HUIs) from databases is an important data mining task, which refers to the discovery of itemsets with high utilities (e.g. high profits). However, it may present too many HUIs to users, which also degrades the efficiency of the mining process. To achieve high efficiency for the mining task and provide a concise mining result to users, we propose a novel framework in this paper for mining closed+high utility itemsets(CHUIs), which serves as a compact and lossless representation of HUIs. We propose three efficient algorithms named AprioriCH (Apriori-based algorithm for mining High utility Closed+itemsets), AprioriHC-D (AprioriHC algorithm with Discarding unpromising and isolated items) and CHUD (Closed+High Utility Itemset Discovery) to find this representation. Further, a method called DAHU (Derive All High Utility Itemsets) is proposed to recover all HUIs from the set of CHUIs without accessing the original database. Results on real and synthetic datasets show that the proposed algorithms are very efficient and that our approaches achieve a massive reduction in the number of HUIs. In addition, when all HUIs can be recovered by DAHU, the combination of CHUD and DAHU outperforms the state-of-the-art algorithms for mining HUIs.

Several techniques have been proposed to automatically grade students' free-text answers in e-learning systems. However, these techniques provide no or limited support for the evaluation of acquired procedural knowledge. To address this issue, we propose a new approach, named ProcMark, specifically designed to assess answers containing procedural knowledge. It requires a teacher to provide the ideal answer as a semantic network (SN) that is used to automatically score learners' answers in plain text. The novelty of our approach resides mainly in three areas: a) the variable granularity levels possible in the SN and the parameterizing of ontology concepts, thus allowing the students free expression of their ideas, b) the new similarity measures of the grading system that give refined numerical scores, c) the language-independence of the grading system as all linguistic information is given as data files or dictionaries and is distinct of the semantic knowledge of the SN. Experimental results in a Computer Algorithms course show that the approach gives marks that are very close to those of human graders, with a very strong (0.70, 0.79, and 0.79) positive correlation.

Mining high utility item sets from transactional databases is an important data mining task, which refers to the discovery of item sets with high utilities (e.g. high profits). Although several studies have been carried out, current methods may present too many high utility item sets for users, which degrades the performance of the mining task in terms of execution and memory efficiency. To achieve high efficiency for the mining task and provide a concise mining result to users, we propose a novel framework in this paper for mining closed+ high utility item sets, which serves as a compact and loss less representation of high utility item sets. We present an efficient algorithm called CHUD (Closed+ High Utility item set Discovery) for mining closed+ high utility item sets. Further, a method called DAHU (Derive All High Utility item sets) is proposed to recover all high utility item sets from the set of closed+ high utility item sets without accessing the original database. Results of experiments on real and synthetic datasets show that CHUD and DAHU are very efficient with a massive reduction (up to 800 times in our experiments) in the number of high utility item sets. In addition, when all high utility item sets are recovered by DAHU, the approach combining CHUD and DAHU also outperforms the state-of-the-art algorithms in mining high utility item sets.

In this article, we propose the Conscious-Emotional Learning Tutoring System technology, a biologically plausible cognitive agent based on human brain functions. This agent is capable of learning and remembering events and any related information such as corresponding procedures, stimuli and their emotional valences. In our model, emotions play a role in the encoding and remembering of events. This allows the agent to improve its behaviour or by remembering previously selected behaviours which were influenced by its emotional mechanism. Moreover, the architecture incorporates a realistic memory consolidation process based on a data mining algorithm.

In previous works, we showed how sequential pattern mining can be used to extract a partial problem space from logged user interactions for a procedural and ill-defined domain where classic domain knowledge acquisition approaches don't work well. In this paper, we describe in details how such a problem space can support important tutoring services such as (1) recognizing the plan of a learner, (2) providing hints and (3) estimating the profile of a learner including its expertise level and missing or misunderstandood skills.

This paper describes how a knowledge model allows training software to evaluate spatial cognitive maps and provide tailored assistance.

In this paper we present an authoring tool milieu that permits modelling graphically any subject-matter domain knowledge and transposing it automatically into related XML files. Generated contents serve as a tutor reasoning support when interacting with students engaged in learning activities through virtual learning environments.