检索结果-内蒙古大学图书馆

Deep Multirepresentation Learning for Data clustering

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2024年第11期35卷 15675-15686页

作者： Sadeghi, Mohammadreza Armanfard, Narges McGill Univ Dept Elect & Comp Engn Montreal PQ H3A 0E9 Canada Mila Quebec AI Inst Montreal PQ H2S 3H1 Canada

Deep clustering incorporates embedding into clustering in order to find a lower-dimensional space suitable for clustering tasks. Conventional deep clustering methods aim to obtain a single global embedding subspace (aka latent space) for all the data clusters. In contrast, in this article, we propose a deep multirepresentation learning (DML) framework for data clustering whereby each difficult-to-cluster data group is associated with its own distinct optimized latent space and all the easy-to-cluster data groups are associated with a general common latent space. Autoencoders (AEs) are employed for generating cluster-specific and general latent spaces. To specialize each AE in its associated data cluster(s), we propose a novel and effective loss function which consists of weighted reconstruction and clustering losses of the data points, where higher weights are assigned to the samples more probable to belong to the corresponding cluster(s). Experimental results on benchmark datasets demonstrate that the proposed DML framework and loss function outperform state-of-the-art clustering approaches. In addition, the results show that the DML method significantly outperforms the SOTA on imbalanced datasets as a result of assigning an individual latent space to the difficult clusters.

关键词： clustering algorithms Training Task analysis Decoding Optimization Learning systems Image reconstruction Autoencoder (AE) cluster-specific AEs data clustering multiple representation learning

来源：评论

学校读者我要写书评

暂无评论

Consensus clustering With Co-Association Matrix Optimization

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2024年第3期35卷 4192-4205页

作者： Shi, Yifan Yu, Zhiwen Chen, C. L. Philip Zeng, Huanqiang Huaqiao Univ Sch Engn Quanzhou 362021 Peoples R China South China Univ Technol Sch Comp Sci & Engn Guangzhou 510006 Peoples R China Pengcheng Lab Shenzhen 518066 Peoples R China

Consensus clustering can derive a more promising and robust clustering result by integrating multiple partitions strategically. However, there are several limitations in the existing approaches: 1) most of the methods compute the ensemble-information matrix heuristically and lack of sufficient optimization;2) the information from the original dataset is rarely considered;and 3) the noise in both label space and feature space is ignored. To address these issues, we proposed a novel consensus clustering method with co-association matrix optimization (CC-CMO), which aims at improving the co-association matrix by taking abundant information from both label space and feature space into consideration. In label space, CC-CMO derives a weighted partition matrix capturing the intercluster correlation and further designs a least squares regression (LSR) model to explore the global structure of data. In feature space, CC-CMO minimizes the reconstruction error with doubly stochastic normalization in the projective subspace to eliminate noise features as well as learn the local affinity of data. To improve the co-association matrix by jointly considering the subspace representation, global structure, and local affinity of data, we explicitly propose a unified optimization framework and design an alternating optimization algorithm for the optimal co-association matrix. Extensive experiments on a variety of real-world datasets demonstrate the superior performance of CC-CMO to the state-of-the-art consensus clustering approaches.

关键词： Optimization clustering algorithms Partitioning algorithms Linear programming Correlation clustering methods Stochastic processes Alternative optimization consensus clustering

来源：评论

学校读者我要写书评

暂无评论

MGNR: A Multi-Granularity Neighbor Relationship and Its Application in KNN Classification and clustering Methods

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2024年第12期46卷 7956-7972页

作者： Xie, Jiang Xiang, Xuexin Xia, Shuyin Jiang, Lian Wang, Guoyin Gao, Xinbo Chongqing Univ Telecommun & Posts Chongqing Key Lab Computat Intelligence Chongqing 400065 Peoples R China

In the real world, data distributions often exhibit multiple granularities. However, the majority of existing neighbor-based machine-learning methods rely on manually setting a single-granularity for neighbor relationships. These methods typically handle each data point using a single-granularity approach, which severely affects their accuracy and efficiency. This paper adopts a dual-pronged approach: it constructs a multi-granularity representation of the data using the granular-ball computing model, thereby boosting the algorithm's time efficiency. It leverages the multi-granularity representation of the data to create tailored, multi-granularity neighborhood relationships for different task scenarios, resulting in improved algorithmic accuracy. The experimental results convincingly demonstrate that the proposed multi-granularity neighbor relationship effectively enhances KNN classification and clustering methods.

关键词： Computational modeling Data models clustering methods Machine learning clustering algorithms Classification algorithms Task analysis clustering granular-ball computing KNN multi-granularity neighbor relationship

来源：评论

学校读者我要写书评

暂无评论

EDMC: Efficient Multi-View clustering via Cluster and Instance Space Learning

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2024年 26卷 5273-5283页

作者： Qin, Yalan Pu, Nan Wu, Hanzhou Shanghai Univ Sch Commun & Informat Engn Shanghai 200444 Peoples R China Univ Trento Dept Informat Engn & Comp Sci I-40128 Trento Italy

Multi-view subspace clustering aims to cluster the data lying in a union of subspaces with low dimensions. The commonly used spectral clustering performs the final clustering based on an n x n affinity graph, which suffers from relative high time and space complexity. Some existing works have chosen key anchors with uniform sampling strategy or K-means for dealing with large-scale datasets. However, few of them pay attention to the physical meaning of cluster representation in the column of the dataset for learning informative anchors, which is independent from the instance representation. In this paper, we propose efficient dual multi-view clustering (EDMC) with relative low complexity. To be specific, EDMC makes full use of cluster representation space in the column of the dataset to help produce informative anchors, which has a clear physical meaning and is independent of instance representation in the row. It simultaneously explores the cluster and instance subspace representations to learn anchors for large-scale datasets. We perform anchor learning and efficient multi-view clustering in a unified framework and then adopt an alternative optimization strategy for solving the formulated problem. Extensive experiments performed on different datasets in terms of several metrics validate the superiority of the proposed method.

关键词： clustering algorithms Tensors Scalability Representation learning Optimization Dimensionality reduction Complexity theory Multi-view subspace clustering cluster representation instance representation anchor learning a unified framework

来源：评论

学校读者我要写书评

暂无评论

Graph Contrastive Learning for clustering of Multi-Layer Networks

引用

IEEE TRANSACTIONS ON BIG DATA 2024年第4期10卷 429-441页

作者： Yang, Yifei Ma, Xiaoke Xidian Univ Sch Comp Sci & Technol Key Lab Smart Human Comp Interact & Wearable Techn Xian 710071 Peoples R China

Multi-layer networks precisely model complex systems in society and nature with various types of interactions, and identifying conserved modules that are well-connected in all layers is of great significance for revealing their structure-function relationships. Current algorithms are criticized for either ignoring the intrinsic relations among various layers, or failing to learn discriminative features. To attack these limitations, a novel graph contrastive learning framework for clustering of multi-layer networks is proposed by joining nonnegative matrix factorization and graph contrastive learning (called jNMF-GCL), where the intrinsic structure and discriminative of features are simultaneously addressed. Specifically, features of vertices are first learned by preserving the conserved structure in multi-layer networks with matrix factorization, and then jNMF-GCL learns an affinity structure of vertices by manipulating features of various layers. To enhance quality of features, contrastive learning is executed by selecting the positive and negative samples from the constructed affinity graph, which significantly improves discriminative of features. Finally, jNMF-GCL incorporates feature learning, construction of affinity graph, contrastive learning and clustering into an overall objective, where global and local structural information are seamlessly fused, providing a more effective way to describe structure of multi-layer networks. Extensive experiments conducted on both artificial and real-world networks have shown the superior performance of jNMF-GCL over state-of-the-art models across various metrics.

关键词： clustering algorithms Network analyzers Representation learning Couplings Complex systems Big Data Analytical models Multi-layer networks graph clustering nonnegative matrix factorization graph contrastive learning

来源：评论

学校读者我要写书评

暂无评论

A Distributed Higher-Order k-Medoids clustering Algorithm for Network Partition

引用

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 2024年第5期11卷 4181-4193页

作者： Mo, Yuanqiu Xing, Ran Hou, Huazhou Southeast Univ Nanjing 211189 Peoples R China Univ Sydney Sydney NSW 2006 Australia Purple Mt Labs Nanjing 211102 Peoples R China

In this paper, we develop a distributed higher-order k-medoids clustering algorithm for networks using hop count as the distance metric, typical examples include social networks, wireless sensor networks and etc. Different than the classical k-medoids clustering where each cluster is assigned with one medoid (representative nodes in the network), the proposed algorithm is capable of partitioning the nodes into clusters dominated by multiple medoids. In this algorithm, a higher-order shortest path algorithm is first adopted to assign each node the clustering set comprising of its h (h >= 1) closest medoids using the hop count metric, then an aggregation algorithm collects the local information from each node to the medoids in its clustering set, based on which a medoids update mechanism helps each medoid decide whether to keep its medoid status or assign the status to one of its neighbors, so as to improve the clustering quality. The clustering performance of the proposed algorithm is proved to monotonically converge to a local optimum after running the algorithm finite times. Simulations results show that the proposed algorithm can achieve superior performance compared with several centralized $k$-medoids clustering algorithms and scale well as the number of nodes increases.

关键词： clustering algorithms Partitioning algorithms Length measurement Wireless sensor networks Social networking (online) Search problems Prototypes k -medoids clustering higher-order clustering graph algorithms

来源：评论

学校读者我要写书评

暂无评论

Energy-Balanced Routing Protocol With Nonuniform clustering for Underwater Acoustic Sensors Networks

引用

IEEE SENSORS JOURNAL 2024年第22期24卷 38082-38091页

作者： Jin, Zhigang Li, Haoyong Wang, Ying Liang, Jiawei Cheng, Simeng Tianjin Univ Sch Elect & Informat Engn Tianjin 300072 Peoples R China

Energy consumption has been the focus of routing protocols in underwater acoustic sensor networks (UASNs), and many cluster-based routing protocols have been proposed to optimize energy consumption. However, there is the "hotspot" problem resulting from frequent data forwarding by cluster heads (CHs) and energy inefficiency caused by the transmission of data packets from shallow water to deep water. Therefore, we propose an energy-balanced routing protocol with nonuniform clustering (ERNC) to balance energy consumption and improve data transmission efficiency. First, without accurate 3-D localization, nodes exchange information with each other, and the combined coordinate of layer ID and hop ID is proposed to represent the node's location information for subsequent CH selection and intercluster routing. Then, the combined coordinate, residual energy, and node density are considered comprehensively to select CHs for making their distribution uneven and equalizing the energy consumption. In the intercluster routing phase, the next hop candidate node sets with different forwarding priorities are constructed based on nodes' coordinates to improve the network transmission efficiency. Moreover, we design the different holding times for the next-hop nodes in the same set to balance the energy consumption of CHs. The simulation results show that ERNC can effectively extend the network lifetime and improve the data transmission performance.

关键词： Energy consumption Routing protocols clustering algorithms Routing Data communication Three-dimensional displays Optimization Underwater acoustics Particle swarm optimization Monitoring Energy consumption balance nonuniform clustering routing protocol underwater acoustic sensor networks (UASNs)

来源：评论

学校读者我要写书评

暂无评论

Design and Hardware Implementation of Low-Latency 4-Splitting Tree-Structure-Based K-Means clustering Trainer and Classifier Chip for Detecting Cybersecurity Attacks

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS 2025年第4期72卷 613-617页

作者： Shih, Xin-Yu Chen, Hsi-Cheng Hung, Xin-Liang MediaTek Dept Wireless Connect Network Hsinchu 300 Taiwan Natl Sun Yat Sen Univ Dept Elect Engn Kaohsiung 804 Taiwan

In this brief, we propose a first system chip architecture of 4-splitting tree-structure-based K-means clustering trainer and classifier. It mainly aims to efficiently detect cybersecurity attacks in the emerging AI-driven world. In addition, we also propose three design techniques to elaborate our chip, featuring a low-latency property in both training and classification aspects. Our developed architecture is first validated with the Xilinx FPGA platform. Furthermore, it is implemented in ASIC and well-verified with five representative cybersecurity-related datasets. The total core area with TSMC 40nm low-Vt CMOS technology only occupies 0.44 um(2), maximally operating at 434.78 MHz and possessing a maximum classification throughput of 3.48 GBps on chip. In comparison, the speed-up ratio of training and classification is up to 9.33 and 8.73 times, respectively, successfully delivering a low-latency characteristic.

关键词： Training Training data Hardware Time division multiplexing Low latency communication Accuracy Machine learning Data mining clustering algorithms Semisupervised learning Tree structure K-means low latency chip architecture cybersecurity detection machine learning semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Fuzzy clustering From Subset-clustering to Fullset-Membership

引用

IEEE TRANSACTIONS ON FUZZY SYSTEMS 2024年第9期32卷 5359-5370页

作者： Chen, Huimin Duan, Yu Nie, Feiping Wang, Rong Li, Xuelong Northwestern Polytech Univ Sch Comp Sci Sch Artificial Intelligence Opt & Elect iOPEN Xian 710072 Peoples R China Northwestern Polytech Univ Key Lab Intelligent Interact & Applicat Minist Ind & Informat Technol Xian 710072 Peoples R China Northwestern Polytech Univ Sch Artificial Intelligence Opt & Elect iOPEN Xian 710072 Peoples R China

Fuzzy theory, which extends precise binary logic to continuous fuzzy logic, provides an effective tool for uncertainty problems in machine learning and thus, evolved fuzzy methods such as fuzzy c-means and fuzzy graph clustering. Among them, graph-based clustering methods have become a hot spot in the field of unsupervised clustering due to their good ability to process nonlinear data. Unfortunately, they usually suffer from high time complexity and cumbersome regularization parameter tuning, so their practical applications are greatly limited. To this end, we propose an efficient graph-cut algorithm called fS2F. Based on the similarity graph between the dataset and landmark subset, fS2F transforms the membership learning of the entire dataset into a clustering problem of representative points, which greatly improves its clustering efficiency. In addition, fS2F softly constrains the cluster size in a way that does not require additional regularization parameters so that it can be widely and conveniently applied. The article also presents the optimization method for this model and demonstrates its effectiveness through experiments.

关键词： clustering algorithms clustering methods Task analysis Eigenvalues and eigenfunctions Tuning Transforms Fuzzy systems Fuzzy clustering graph cut label transmission parameter-free

来源：评论

学校读者我要写书评

暂无评论

DGBPSO-DBSCAN: An Optimized clustering Technique Based on Supervised/Unsupervised Text Representation

引用

IEEE ACCESS 2024年 12卷 110798-110812页

作者： Abdulsahib, Asma Khazaal Balafar, M. A. Baradarani, Aryaz Univ Tabriz Fac Elect & Comp Engn Tabriz *** East Azerbaijan Iran Tessonics Inc Ctr Diagnost Imaging Res Windsor ON N9A Canada

Density-based spatial clustering of noisy applications (DBSCAN), a widely used density-based clustering technique, faces challenges in determining its key parameter, Eps, leading to manual specification and suboptimal clustering outcomes. Additionally, its time complexity poses limitations. This study introduces a novel approach to enhance the algorithm's performance. We combined the DBSCAN algorithm with another approach, the particle swarm optimization, based on a novel way to represent text, termed a dependency graph;this method is recognized as DGBPSO-DBSCAN (Dependency Graph Based Particle swarm algorithm for DBSCAN). This paper focuses on employing a novel approach for PSO variable upgrading to explore the Eps range more rapidly and effectively. This method offers a selection of informative elements from the text, where the initial groups are derived from the graph-based degree centrality, the PSO method is used to choose the most compelling features, and the DBSCAN algorithm is used for text clustering. The experimental findings indicate that the modified PSO is used to improve DBSCAN. We compared the outcomes of the suggested technique to those of the standard clustering algorithm. According to the assessment criteria MSE, accuracy, precision, recall, and F-measure, our technique has been demonstrated to be superior to the conventional method.

关键词： clustering algorithms Reviews Particle swarm optimization Accuracy Partitioning algorithms Optimization Noise measurement Decision trees DBSCAN algorithm decision tree degree centrality dependency graph

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：