Search

• Solve the imbalance problem in cardiovascular disease data. • A novel method named ADWGKFO is proposed to transform the unbalanced data sets. • A more targeted sampling weight determination method is proposed. • A sample testing method based on Gaussian kernel is proposed to filter new samples. • Empirical analysis shows the proposed method performs better than similar methods. In risk assessment of cardiovascular disease (CVD), the classification error caused by unbalanced data is a significant challenge, which has sparked widespread concern and research upsurge in the field of data mining. Therefore, in view of the imbalance of CVD data sets, an oversampling method via adaptive double weights and Gaussian kernel function (ADWGKFO) is proposed, which converts the unbalanced data sets into balanced data sets. Firstly, clustering algorithm is utilized to cluster minority samples, boundary samples are identified by Borderline-Synthetic Minority Over-sampling Technique (Borderline-SMOTE), K nearest neighbor and support vector machine algorithms, and the number of samples synthesized in each group is calculated based on the double weights of boundary points and majority distribution. Secondly, in order to clearly define the classification boundary, the mutual class potential of new samples in each cluster is calculated by Gaussian kernel function, and new samples are filtered according to the mutual class potential until the data set is balanced. Finally, taking the data sets from Kaggle platform as the research samples, the proposed method is empirically analyzed. In order to validate the efficacy and universality of the proposed method, this paper selects CatBoost that is a new integrated algorithm to test the effect of the ADWGKFO method, and compares it with different sampling methods and different classifiers using performance evaluation indexes such as accuracy, F1-score and area under the curve (AUC). Compared with the combinations of other methods, the accuracy, F1-score and AUC are significantly improved. It is concluded that the ADWGKFO method described in this paper can successfully improve the data quality, and increases the reliability of CVD risk assessment. [ABSTRACT FROM AUTHOR]

In this paper, we discuss a rough set approach to missing attribute values. Among many ways of interpreting missing values, in this paper we focus on two interpretations, lost values and "do not care" conditions. Using these interpretations, global and saturated probabilistic approximations are constructed with two types of granules: characteristic sets and maximal consistent blocks. We compare eight approaches, combining two interpretations of missing attribute values, two types of probabilistic approximations with two types of granules using an error rate that is computed as a result of ten-fold cross-validation. Using a 5% level of statistical significance, we present the experimental results for these eight approaches, showing statistically significant differences between all approaches to mining incomplete data. The results also show that no one method and approach is the best for every data set and that all eight approaches should be attempted. The final section of the paper presents the idea of concept-compatible data sets. We show that for these types of data sets, global and saturated probabilistic approximations for a concept are identical to the concept. We also show that for an incomplete data set with no duplicate rows using the lost interpretation of missing attribute values, the data set is concept-compatible. • Two interpretations of missing attribute values: lost values and "do not care" conditions are considered • Global and saturated probabilistic approximations are constructed from characteristic sets and maximal consistent blocks • Eight approaches to mining incomplete data sets are compared and significant differences between them are indicated • A novel idea of the concept-compatible data sets is introduced [ABSTRACT FROM AUTHOR]

Multilabel classification is a growing paradigm in the fields of data mining and machine learning. In multilabel learning, each input sample can belong to more than one binary class (termed "label") in a multilabel framework, with contrasts with the standard single-label learning scenario. The challenge of producing a better classifier lies in the advantageous use of the correlations among different labels. In recent years, many multilabel models have been proposed, making it difficult to decide which methods to use. In this paper, we present the most comprehensive ranking performance comparison carried out thus far among numerous methods. We conduct a comprehensive analysis by comparing several configurations of 56 distinct methods, resulting in a total of 173 trained models. In addition, we utilize an extensive collection of problems involving 65 datasets. Furthermore, we analyze the effectiveness of the tested techniques by evaluating their performance with six different ranking performance metrics. Our findings indicate that while certain strategies consistently rank highly among the top-performing models, the most effective methods are strongly correlated with the specific metrics used to assess their performance. Finally, we examine many behavioral aspects of the different approaches. [ABSTRACT FROM AUTHOR]

The Multiparty Private Set Intersection (MPSI) protocol allows untrusted parties to compute the intersection of their sets without revealing any additional information. MPSI has found wide applications in fields such as advertising, vertical federated learning, privacy-preserving data mining, and private contact tracing. However, existing MPSI solutions face challenges in terms of security and efficiency. Some MPSI schemes are secure only against semi-honest adversaries and cannot effectively defend against malicious adversaries. Other schemes, while capable of withstanding malicious adversaries, exhibit efficiency problems, especially as the number of corrupt participants increases. A practical MPSI solution must maintain high security and efficiency, even when handling large-scale data and a majority of corrupted participants. This paper introduces a novel MPSI scheme named TH-PSI, built upon a new star-shaped topological network communication framework based on the trusted execution environment. TH-PSI has been rigorously proven to be secure within the standard Universal Composability framework, capable of thwarting malicious adversaries. To demonstrate the efficiency of TH-PSI, we conduct a comprehensive series of experiments. We implement the TH-PSI scheme and tested it in both LAN and WAN settings, involving up to 30 participants and datasets as large as 222. We compare TH-PSI with state-of-the-art MPSI schemes, ENOC21 and NTY21, with a specific focus on their efficiency in the presence of malicious adversaries. Our experiments show that TH-PSI is an efficient solution capable of delivering results within a matter of seconds, even when dealing with data at the million-level and a majority of corrupted participants. [ABSTRACT FROM AUTHOR]

Time series clustering serves as a potent data mining method, facilitating the analysis of an extensive array of time series data without the prerequisite of any prior knowledge. It finds wide-ranging use across various sectors, including but not limited to, financial and medical data analysis, and sensor data processing. Given the high dimensionality, non-linearity, and redundancy characteristics associated with time series, conventional clustering algorithms frequently fall short in yielding satisfactory results when directly applied to this kind of data. As such, there is a critical need to judiciously select suitable feature extraction methods and dimension reduction techniques. This paper introduces a time series clustering algorithm, drawing primarily from polynomial fitting derivative features as a wellspring for feature extraction to achieve effective clustering results. Initially, Hodrick Prescott (HP) filtering comes into play for the processing of raw time series data, thereby eliminating noise and redundancy. Subsequently, polynomial curve fitting (PCF) is applied to the data to derive a globally continuous function fitting this time series. Next, by securing multi-order derivative values via this function, the time series is transformed into a multi-order derivative feature sequence. Lastly, we designed a polynomial function derivative features-based dynamic time warping (PFD_DTW) algorithm for determining the distance between two equal or unequal granular length time series, and subsequently a hierarchical clustering method anchored on the PFD_DTW distances for time series clustering after computing interspecies distances. The effectiveness of this method is corroborated by experimental results obtained from several practical datasets. [ABSTRACT FROM AUTHOR]

Related to many applications in different fields, such as game theory, information fusion, data mining, and decision making, we have introduced in one our previous paper so called generalized Choquet-type integral for a real-valued function concerning a set-function and a σ -additive measure. The present study further generalizes the generalized Choquet-type integral in terms of a double set-function Choquet integral for a real-valued function based on a set-function and fuzzy measure. Several of its properties and convergence theorems are obtained, and a novel type of Jensen's inequality is proved. The stability of the proposed system formed by a double set-function Choquet integral concerning multiple inputs and one output is indicated. An effective application in decision making is shown through numerical examples. [ABSTRACT FROM AUTHOR]

Fuzzy inference with the Bandler-Kohout subproduct (BKS) has been successfully applied in many fields such as fuzzy control, artificial intelligence, image processing, data mining, decision-making, prediction, classification and so on. However, one has to face with the rule explosion in these applications. To deal with this problem, hierarchical fuzzy systems with the compositional rule of inference (CRI) method have been constructed by a series of low-dimensional sub fuzzy systems. And it has been proved that hierarchical fuzzy inference method can efficiently restrain the explosion of fuzzy rules. Therefore, in order to increase the computational efficiency of the fuzzy inference based on the BKS when multi-input-single-output (MISO) fuzzy rules are involved, this paper mainly constructs two hierarchical fuzzy inference methods based on the BKS in which the if-then rules are respectively interpreted by fuzzy implications and ML -implications. Moreover, the validity of the two BKS hierarchical fuzzy inferences is studied with the GMP rules. Finally, two examples are employed to illustrate the computational efficiency of our proposed BKS hierarchical inference methods. [ABSTRACT FROM AUTHOR]

Today, more and more data is collected and stored by different organizations. When the data is distributed among multiple users who wish to perform data mining on the joint data, outsourcing the task to cloud servers becomes an attractive solution for users who lack professional skills. However, privacy concerns make users reluctant to adopt such solutions. In this paper, we propose a privacy-preserving outsourced k -means clustering (PPOKC) algorithm on distributed data. Our system consists of multiple users and two non-colluding servers. Each user submits their data to the servers, which perform k -means clustering over the joint data without compromising data privacy. Unlike existing solutions, which typically include a cryptographic server and a computing server, each server in our system provides both services. Based on this architecture, we first design several important sub-protocols, including secure comparison, secure minimum and secure division. We then use these protocols to construct an efficient and highly secure PPOKC algorithm. The implementation of our algorithm relies heavily on secret sharing techniques, complemented by homomorphic encryption. We also conduct theoretical analysis and experiments. Security analysis shows that our scheme guarantees the security of the input/output data, the intermediate results and the data access pattern under the semi-honest model. Complexity analysis and numerical experiments show that our algorithm has good efficiency and is suitable for practical applications. [ABSTRACT FROM AUTHOR]

Clustering, as an essential technique in unsupervised learning, plays a pivotal role in the fields of data mining and machine learning. However, the classic K -means clustering algorithm has intrinsic drawbacks such as sensitivity to initial cluster centers, susceptibility to a local optimal solution, and challenges in handling data uncertainty. To address these problems, this paper proposes an artificial hummingbird algorithm (AHA)-based three-way K -means clustering algorithm, called AHA-3WKM. First, AHA is introduced to address the problems of sensitivity to initial cluster centers and local optima. Second, a fitness function of AHA is specifically constructed to find the best initial clustering centers so that the hummingbirds can search for high-quality food sources, i.e., the global optimum cluster centers. Third, a three-way clustering approach is utilized to capture information about data uncertainty. In this way, the results of clustering are divided into three distinct regions based on the relationship between objects and clusters. The experimental results demonstrate that AHA-3WKM has good performance, and enhances the stability and the accuracy of clustering results. • AHA is introduced to address the problems of the sensitivity to initial cluster centers and the proneness to local optima. Hummingbirds are treated as data points, which dynamically update their strategies and effectively find the optimal cluster centers during multiple iterations. • A fitness function is designed based on the clustering principle of "birds of a feather flock together", with the aim of simplifying calculations, which enhances the specificity and practicality of K-means algorithm. • An AHA-based three-way K-means clustering algorithm (i.e., AHA-3WKM) is proposed. The clustering process is initialized with cluster centers optimized by AHA, and the results are represented in three regions, which can capture the uncertainty within the datasets. [ABSTRACT FROM AUTHOR]

Feature selection has become a mandatory task in data mining, due to the overwhelming amount of features in Big Data problems. To handle this high-dimensional data and avoid the well-known curse of dimensionality, we need to pre-select an optimal subset of features to reduce redundant computations. Federated learning is a machine learning technique based on training an algorithm over many decentralized edge devices holding local rather than global data on a centralized server. Application of this technique is extending to fields such as self-driving cars, medicine and health, and Industry 4.0, where data privacy is compulsory. Feature selection through federated learning is a complicated task since suboptimal features calculated by feature selection methods may be different in heterogeneous datasets from different nodes. In this paper, we propose a lossless federated version of the classic minimum redundancy maximum relevance (mRMR) feature selection algorithm, called federated mRMR (fed-mRMR), which, without losing any effectiveness of the original mRMR method, is applicable to federated learning approaches and capable of dealing with data that are not independent and identically distributed (non-IID data). Implementation can be found at: https://github.com/jorgehermo9/fed-mrmr • Feature Selection (FS) is a key step in machine learning. • Classical FS does not consider privacy preservation. • Federated FS allows to perform feature selection in federated environments. • fed-mRMR is a federated feature selection tool publicly available. • fed-mRMR performs better than original's mRMR proposal in terms of computation. [ABSTRACT FROM AUTHOR]

This paper introduces a novel machine learning approach, called BBATDD (Boosting-Based Algorithm Trained with Drift Detector), that departs from traditional epoch-based methods by adopting a data stream generation strategy from the original dataset. The proposed method is versatile, applicable to a broad range of deep structures and other machine learning architectures. The key to its effectiveness is the conversion of original data into a streaming sequence, which is then fed into the neural network structure. To monitor the learning process and prevent overfitting, a drift detector, commonly used in stream data mining, is employed. Additionally, two methods, ELB and ENT, are introduced to prioritize problematic data elements during the learning process, enhancing overall model performance. Extensive simulations on benchmark datasets (MNIST and CIFAR) validate that BBATDD achieves higher accuracy with significantly fewer batches, without sacrificing generalization capability. For instance, when applied to the MNIST dataset and processing 10 000 batches while comparing loss function values, our approach demonstrates a remarkable 45.6% improvement. [ABSTRACT FROM AUTHOR]

Heterogeneous graphs (HGs) contain many nodes and their interaction relationships, which can model complex systems and provide rich semantic and structural information for task execution. Among these, HG representation stands as the fundamental and pivotal component. Existing HG representation methods primarily employ graph neural networks to acquire the semantics of nodes along various meta-paths and fuse them to represent the nodes. The most prevalent HG representation methods encompass two steps: semantic information extraction within meta-paths and semantic fusion between meta-paths. However, these methods overlooked the consideration of node heterogeneity within meta-paths and the simultaneous semantic correlation between meta-paths. Specifically, node heterogeneity within meta-paths signifies that the meta-path-based neighbors do not consistently contain information that positively influences the target node, and the semantic correlation between meta-paths indicates that different meta-path spaces are not entirely independent. Disregarding either of these issues leads to the propagation of irrelevant or redundant information and potential disruption of HG embedding. Consequently, in this study, we propose the HBHG, which is a hierarchical bottleneck for heterogeneous graph representation. HBHG primarily employs the information bottleneck (IB) as a guiding principle, constraining the propagation of irrelevant information within and between meta-paths while preserving relevant information. The central concept of the IB revolves around viewing model learning as the preservation of relevant information and compression of irrelevant information, accomplished by minimizing the dependency between input and hidden features through mutual information (MI) and maximizing the dependency between hidden features and ground-truth. Considering the complexity associated with MI estimation, this paper introduces a novel dependency index, namely the Hilbert-Schmidt independence criterion (HSIC), which offers ease of calculation. Specifically, HBHG comprises two primary components: a semantic bottleneck within meta-paths and a semantic bottleneck between meta-paths. The semantic bottleneck within meta-paths relies primarily on the HSIC-based limitations of dependencies at different layers of the graph neural network on various meta-paths, thereby maximizing the extraction of information relevant to the target node from neighboring nodes. The semantic bottleneck between meta-paths enables flexible extraction and fusion of semantic information based on downstream tasks, achieved by managing the trade-off of dependencies with HSIC between different meta-path semantic spaces. In summary, the proposed HBHG integrates hierarchical bottleneck constraints within and between meta-paths. This integration serves to maximize the aggregation of relevant information while effectively compressing irrelevant information, thereby enhancing the quality of heterogeneous graph embedding. The effectiveness of HBHG was validated through performance and ablation experiments conducted on multiple datasets. [ABSTRACT FROM AUTHOR]

With the advancement of data technology and storage services, the scale and complexity of data are rapidly growing. Consequently, promptly analyzing data and deriving precise insights have become urgent. Nevertheless, traditional methods struggle to balance the speed and accuracy of data mining. This paper proposes a data analysis technique called the Division-Mining-Fusion (DMF) strategy to tackle this challenge. Specifically, we divide a large-scale and complex dataset into multiple small-scale and simple sub-datasets. Then, we extract the knowledge embedded within each sub-dataset. Finally, we combine the extracted knowledge from each sub-dataset to accomplish learning tasks. To demonstrate the superior performance of the DMF strategy, we apply it to two fields: rough set theory and feature selection. The DMF strategy can accelerate the speed of data mining, enhance the accuracy of data analysis, and reduce the dimensionality of data. These advantages suggest that the DMF strategy outperforms traditional methods in processing data more efficiently. In addition, the number of sub-datasets is a crucial parameter of the DMF strategy. As the number of sub-datasets increases, the ability of the DMF strategy to analyze data continuously improves. [ABSTRACT FROM AUTHOR]

This paper presents a study on the location problem of single-server battery swap stations, identifying instances of excessively long waiting times at certain stations during their operation in a real-world company scenario. This study innovatively transforms the problem into an extended version of the classic maximal covering location problem, incorporating technology selection and three sets of additional constraints: budget, closest-assignment, and average waiting time constraints. In particular, the closest-assignment constraint is specifically designed to mimic human behavioral patterns. A non-convex integer programming model is proposed and then reformulated thus enabling it to be tackled by some general-purpose solvers. The model is validated considering a test bed of randomly generated instances. The model comprehensiveness added-value is assessed. Finally, real-world data is used which allows discussing the extraction of information in the studied context. That information is used to build an instance that is solved and analyzed to provide valuable managerial insights, emphasizing the importance of technological improvement to the management. [ABSTRACT FROM AUTHOR]

Multi-view clustering has become one of the popular clustering branches with data accumulation from multiple domains. Unfortunately, a large heterogeneity gap exists due to cross-view discrepancy, resulting in inaccurate sample-wise similarity estimation. Furthermore, the popular multi-view clustering methods heavily rely on sample-wise similarity measurements, which often lead to suboptimal clustering performance due to inaccurate similarity estimation across views. To tackle these challenges, this paper presents an accurate multi-view clustering method from a standpoint of inter-view feature-wise matching to bypass the inaccurate sample-wise similarity that leverages tensor feature matching for cross-viewed yet uniform sample assignment, named by multi-view clustering via tensor feature matching (MC-TFM). Specifically, our approach begins with exploring a comprehensive feature matching tensor by exploiting both intra-view and inter-view correlations among multi-view features. Subsequently, a feature matching matrix, which preserves the correlations and importance of multi-view features for guiding the cross-viewed yet uniform sample assignment, is estimated based on this tensor. Furthermore, these correlations and importance are maintained into two coded feature bases by decomposing the feature matching matrix. Finally, a sample assignment matrix is learned via jointly reconstructing the samples using the two bases further cooperating with spectral clustering. In this way, the heterogeneity gap is bridged by tensor feature matching, and the inaccurate sample-wise similarity estimation is omitted by using feature-wise matching to guide sample-wise assignment. Extensive experiments conducted on seven real-world datasets highlight the effectiveness of the cross-viewed yet uniform sample assignment, demonstrating the potential of our approach in accurate multi-view clustering tasks. • MC-TFM leverages inter-view feature-wise matching to bypass the inaccurate sample-wise similarity estimation. • MC-TFM utilizes the feature matching tensor for sample assignment to exploit both intra-view and inter-view correlations. • Demonstration of the substantial potential of MC-TFM through extensive experiment. [ABSTRACT FROM AUTHOR]

In recent years, rare pattern mining has shown great vitality in some real-world fields, such as disease diagnosis, criminal behavior analysis, anomaly detection in networks, and so on. When data organizations publish or share information publicly, shared data can be at risk of leakage as data mining techniques may discover sensitive knowledge and information. To keep competitors from obtaining hidden information after processing the database, privacy-preserving data mining (PPDM) has been proposed and studied widely. However, most of the techniques in PPDM are applied to frequent pattern mining and cannot deal with the privacy protection problems in rare pattern mining, such as network vulnerability detection and abnormal medical data. To address this limitation, we introduce a privacy-preserving technique for rare pattern mining. In this paper, two novel algorithms named Longest Transaction-Minimum Item Number (LT-MIN) and Longest Transaction-Maximum Item Number (LT-MAX) are proposed to hide sensitive rare itemsets and return the sanitized database. These two algorithms succeed in hiding target itemsets while minimizing the side effects on the original database. What's more, they employ a projection mechanism to reduce the time spent scanning the database. Besides using the traditional evaluation criteria in PPDM, we also propose two additional similarity measures to evaluate the performance from the perspective of the itemsets and the structural integrity of the database. The experimental results indicate that the proposed algorithms can hide sensitive rare itemsets successfully and efficiently, and the evaluation methods used can become the evaluation criteria for privacy-preserving rare itemset mining (PPRIM). [ABSTRACT FROM AUTHOR]

• MRIformer is designed based on the characteristics of wind speed data. • ASI block is used to realize subsequence information mining and interaction. • MRI block is used to realize multi-resolution data mining and information fusion. • The proposed model is compared with twelve mainstream prediction models. Wind speed prediction is crucial for managing energy consumption in wind farms. Traditional wind speed prediction techniques often overlook two essential characteristics of wind speed data: (a) the downsampled wind speed data can retain cyclic and trend information, which is valuable for the model. (b) Multi-resolution speed data exhibited distinct patterns, enabling the model to extract insights from various perspectives. Considering the above two characteristics, this paper presents a novel approach called the Multi-Resolution Interactive transformer (MRIformer), which consists of the ASI block and the MRI block. The ASI block utilizes two different attention mechanisms to extract temporal information and enhance interactive learning among subsequences while downsampling wind speed data. The MRI block utilizes a tree structure to stack multiple layers of ASI blocks, enabling the analysis of wind speed data at various resolutions. By incorporating residual connections and multi-head attention, the MRI block effectively fuses data with different resolutions. Comparative experiments on three real-world datasets led to the following conclusions. (a) MRIformer exceeded 14 state-of-the-art baselines on all datasets and achieved a performance improvement of over 7.5%. (b) The effectiveness of the designed structure is demonstrated through component replacement and ablation experiments. [ABSTRACT FROM AUTHOR]

Multivariate time series classification is a crucial task with applications in broad areas such as finance, medicine, and engineering. Transformer is promising for time series classification, but as a generic approach, they have limited capability to effectively capture the distinctive characteristics inherent in time series data and adapt to diverse architectural requirements. This paper proposes a novel dynamic transformer-based architecture called Dyformer to address the above limitations of traditional transformers in multivariate time series classification. Dyformer incorporates hierarchical pooling to decompose time series into subsequences with different frequency components. Then, it employs Dyformer modules to achieve adaptive learning strategies for different frequency components based on a dynamic architecture. Furthermore, we introduce feature-map-wise attention mechanisms to capture multi-scale temporal dependencies and a joint loss function to facilitate model training. To evaluate the performance of Dyformer, we conducted extensive experiments using 30 benchmark datasets. The results unequivocally demonstrate that our model consistently outperforms a multitude of state-of-the-art methods and baseline approaches. Our model also copes well with limited training samples when pre-trained. • We present a transformer-based dynamic architecture to achieve adaptive learning strategies for different frequency components of the time series data. • We design a hierarchical pooling layer to decompose time series into subsequences representing different frequency components to facilitate time series classification. • We implement two types of gates for allowing the architecture to dynamically adapt the layer activation to realize a flexible structure. [ABSTRACT FROM AUTHOR]

In recent years, sequential pattern mining (SPM) has been applied in many domains, including recommender systems, fraud detection, and other related industries. Target-oriented SPM (TSPM) has been proposed to improve the interpretability of SPM by addressing the issue that too many irrelevant or meaningless sequential patterns are often discovered, which facilitates data analysis. Nevertheless, current TSPM research does not consider contiguity to identify sequential patterns, despite the fact that it is an important feature of sequence data. Contiguity has significant implications in applications such as bioinformatics and network intrusion detection, as it can aid the detection of periodic patterns and anomalous behaviors. Therefore, this paper proposes a novel problem of target-oriented contiguous sequential pattern mining (TCSPM) and introduces two algorithms, called TCSPM and TCSPM+. TCSPM utilizes a compact data structure called sequence chain and relies on a sequence chain pruning strategy to reduce the search space. In addition, TCSPM applies a reverse matching technique to quickly search for targeted contiguous sequences. TCSPM+ is an optimized version of TCSPM that utilizes sequence segmentation operations and a query sequence pruning strategy to further improve performance. Extensive experiments on both real and synthetic datasets demonstrate that TCSPM is an efficient algorithm for discovering targeted contiguous sequential patterns (TCSPs). TCSPM+ substantially enhances the performance of TCSPM in terms of runtime, memory usage, and scalability, outperforming TCSPM by an order of magnitude on some datasets. [ABSTRACT FROM AUTHOR]