检索结果-内蒙古大学图书馆

A Novel Fuzzy clustering Method Based on Topological Connected Component

IEEE TRANSACTIONS ON FUZZY SYSTEMS 2024年第3期32卷 1050-1062页

作者： Zheng, Bowen Ma, Liwen Beijing Univ Posts & Telecommun Sch Sci Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun Key Lab Math & Informat Networks Minist Educ Beijing 100876 Peoples R China

With the development of fuzzy set theory, the clustering methods widely used in various fields can be divided into hard clustering methods and fuzzy clustering methods. Fuzzy clustering methods use the membership function with values in the interval [0, 1] to describe fuzzy concepts, which is different from the idea of "non-0 is 1" in hard clustering methods, so they are more suitable for some real datasets. In this article, we introduce a novel fuzzy clustering method called fuzzy connected clustering (FCC), which is based on the idea of topological connected components. According to the basic view of topology, FCC directly considers the location characteristics between objects, unlike the common clustering methods that need artificial conditions, such as distance scale, cluster center, fuzzy similar matrix, etc. We use topological knowledge to derive FCC through a series of strict logical reasoning, and design a new algorithm, FCC algorithm. Furthermore, we propose a method of constructing "partial order" to express the clustering results of FCC intuitively. Combining with the application examples of FCC, we not only demonstrate the FCC's high interpretability in dealing with real-world problems, but also compare FCC with the commonly used fuzzy clustering method FEMCM. Moreover, we also try to use a benchmark dataset to show how well the FCC deals with real problems related to fuzzy concepts. Finally, we further discuss the extensions and limitations of FCC, and give some basic directions to effectively address these limitations, as well as prospects for future work related to this novel clustering method FCC.

关键词： clustering methods FCC clustering algorithms Topology Telecommunications Linear programming Fuzzy set theory Connected component fuzzy clustering fuzzy component clustering (FCC) partial order topology

来源：评论

学校读者我要写书评

暂无评论

Interrelate Training and clustering for Online Speaker Diarization

引用

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING 2024年 32卷 1352-1364页

作者： Chen, Yifan Cheng, Gaofeng Yang, Runyan Zhang, Pengyuan Yan, Yonghong Chinese Acad Sci Inst Acoust Key Lab Speech Acoust & Content Understanding Beijing 100190 Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China

In clustering-based speaker diarization systems, the embedding clusters for distinctive speakers exhibit wide variability in size and density, posing difficulty for clustering accuracy. In spite of this, with the assistance of the overall distance relationships among speaker embeddings, most of the embeddings can be grouped to the correct cluster by sophisticated offline clustering algorithms. However, in online scenarios, such a complete distance relationships of the embeddings can not be obtained due to the incremental arrival of embeddings. Consequently, determining the number of clusters and then correctly grouping the embeddings become challenging in an online fashion. Furthermore, errors would accumulate quickly over time if the online clustering algorithm assigns the embeddings into clusters erroneously in the beginning. To address these problems, we designed a novel framework for online clustering. To reduce the high variability of speaker embeddings, we proposed the clustering guided embedding extractor training (CGEET) algorithm to encourage similarity between the size of the embedding space for different speakers in attempt to simplify the distance relationships of embeddings. The CGEET algorithm can grasp the distance information of the entire speaker embedding space and provide it to the online clustering algorithm. With this preliminary information, the distance thresholds guided online clustering (DTGOC) algorithm then processes incoming embeddings using a divide-and-conquer approach. It first handles the embeddings with explicit distance relationships and then searches for possible path combination they have with remaining embeddings in an online fashion. Moreover, in order to utilize the distance relationships of embeddings that are far apart in time, an online re-clustering strategy is incorporated in our DTGOC algorithm, which can alleviate error accumulation during online clustering. By implementing the above innovations, our proposed online clus

关键词： clustering algorithms Training Task analysis Oral communication Speech processing Acoustics Machine learning algorithms Metric embedding learning online clustering speaker diarization

来源：评论

学校读者我要写书评

暂无评论

Graph-Regularized Non-Negative Matrix Factorization for Single-Cell clustering in scRNA-Seq Data

引用

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 2024年第8期28卷 4986-4994页

作者： Jiang, Hanjing Wang, Mei-Neng Huang, Yu-An Huang, Yabing Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Key Lab Image Informat Proc & Intelligent Control Minist China Wuhan Peoples R China China Univ Min & Technol Sch Comp Sci & Technol Xuzhou 221116 Peoples R China Yichun Univ Sch Math & Comp Sci Yichun 336000 Peoples R China Northwestern Polytech Univ Sch Comp Sci Xian 710129 Peoples R China Wuhan Univ Renmin Hosp Dept Pathol Wuhan 430060 Peoples R China

The advent of single-cell RNA sequencing (scRNA-seq) has brought forth fresh perspectives on intricate biological processes, revealing the nuances and divergences present among distinct cells. Accurate single-cell analysis is a crucial prerequisite for in-depth investigation into the underlying mechanisms of heterogeneity. Due to various technical noises, like the impact of dropout values, scRNA-seq data remains challenging to interpret. In this work, we propose an unsupervised learning framework for scRNA-seq data analysis (aka Sc-GNNMF). Based on the non-negativity and sparsity of scRNA-seq data, we propose employing graph-regularized non-negative matrix factorization (GNNMF) algorithm for the analysis of scRNA-seq data, which involves estimating cell-cell sparse similarity and gene-gene sparse similarity through Laplacian kernels and p-nearest neighbor graphs (p-NNG). By assuming intrinsic geometric local invariance, we use a weighted p-nearest known neighbors (p-NKN) to optimize the scRNA-seq data. The optimized scRNA-seq data then participates in the matrix decomposition process, promoting the closeness of cells with similar types in cell-gene data space and determining a more suitable embedding space for clustering. Sc-GNNMF demonstrates superior performance compared to other methods and maintains satisfactory compatibility and robustness, as evidenced by experiments on 11 real scRNA-seq datasets. Furthermore, Sc-GNNMF yields excellent results in clustering tasks, extracting useful gene markers, and pseudo-temporal analysis.

关键词： Kernel Laplace equations Bioinformatics clustering algorithms Sparse matrices Data mining Unsupervised learning scRNA-seq non-negative matrix factorization clustering gene marker

来源：评论

学校读者我要写书评

暂无评论

Integrating Curriculum Learning With k-Means: A Data-Centric Approach to Faster clustering

引用

IT PROFESSIONAL 2024年第5期26卷 36-46页

作者： Majeed, Abdul Hwang, Seong Oun Gachon Univ Dept Comp Engn Seongnam 13120 South Korea

k-means clustering is a very popular method that groups n observations into k clusters based on the nearest mean, whereas each observation serves as the prototype of a cluster. Although k-means yields desirable results in most cases, the computing overhead is very high even for modest-size datasets, which makes it unsuitable for larger datasets. To lower computing overhead without degrading performance, in this article, we propose and implement a curriculum learning (CL)-integrated k-means clustering method for efficiently clustering observations. In our method, the data to be clustered via k-means are evaluated beforehand and sorted based on complexity using the CL approach. By applying CL, the data portion that is naturally clustered is identified and bypassed by k-means so that only some complex portions of the data undergo processing, leading to a significant reduction in computing overhead.

关键词： Databases clustering Methods Prototypes clustering algorithms Benchmark Testing Complexity Theory Artificial Intelligence Curriculum Learning Computation Time Big Data Data Mining Machine Learning Techniques Unsupervised Learning Uneven Distribution Data clustering Partial Data Benchmark Datasets Cluster Centers K Means Algorithm Portion Of Data clustering Process High Coherence Knowledge Discovery Choice Scenarios Clusters Of Data Points Unnecessary Computation Results In Most Cases Dominant Values Davies Bouldin Index Real Life Datasets Clusters Of Points Feature Classification Dataset Characteristics Size Classes Numerous Features Distance Calculation clustering Effect

来源：评论

学校读者我要写书评

暂无评论

Relative Comparison-Based Consensus Learning for Multi-View Subspace clustering

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2024年第12期34卷 12376-12387页

作者： Xiao, Xiaolin Wu, Yue Gong, Yue-Jiao South China Normal Univ Sch Comp Sci Guangzhou 510631 Peoples R China South China Univ Technol Sch Comp Sci & Engn Guangzhou 510006 Peoples R China Univ Elect Sci & Technol China Shenzhen Inst Adv Study Shenzhen 610056 Peoples R China

Current multi-view subspace clustering methods typically consist of a within-view module, which explores inherent characteristics using the self-expressive coefficient matrix, and a cross-view module, which promotes consensus among all views toward similar strengths. However, the self-expressive coefficients are directly influenced by the characteristics and distributions of input features, and coefficient matrices with varying strengths may indicate the same clustering structure. Therefore, directly regularizing the coefficient matrices towards a common matrix is unnecessary and may even diminish the clustering performance. We find that it is the relative data relationship, rather than the absolute similarity, that plays a pivotal role in clustering. Building on this realization, we propose a relative comparison measure that enables a more contextual understanding of the data relationship. Subsequently, we develop a Relative Comparison-based Consensus Learning (RCCL) model for multi-view subspace clustering, which encourages the relative data similarities to be consistent across different views. Our RCCL model advances in identifying the underlying data relationship, avoiding unnecessary constraints on absolute consistency, and thereby delving into the fundamental nature of multi-view consensus. We introduce an elegant transformation operator for relative comparison and solve RCCL under the framework of alternating direction method of multipliers. Extensive experiments unequivocally demonstrated the superiority of RCCL.

关键词： clustering algorithms Data models Kernel Computational modeling Vectors Tensors Symmetric matrices Consensus learning low-rank representation multi-view subspace clustering relative comparison

来源：评论

学校读者我要写书评

暂无评论

CBCG: A clustering Algorithm Based on Bidirectional Conical Information Granularity

引用

IEEE TRANSACTIONS ON FUZZY SYSTEMS 2024年第8期32卷 4388-4400页

作者： Yu, Bin Zheng, Zijian Cai, Mingjie Pedrycz, Witold Xu, Zeshui Hunan Normal Univ Coll Informat Sci & Engn Changsha 410081 Peoples R China Hunan Univ Coll Math Changsha 410082 Peoples R China Hunan Univ Shenzhen Res Inst Shenzhen 518100 Peoples R China Univ Alberta Dept Elect & Comp Engn Edmonton AB T6G 2R3 Canada Polish Acad Sci Syst Res Inst PL-00901 Warsaw Poland Istinye Univ Fac Engn & Nat Sci Dept Comp Engn TR-34010 Sariyer Istanbul Turkiye Sichuan Univ Business Sch Chengdu 610064 Peoples R China

In this article, we propose a novel center-based clustering algorithm based on bidirectional conical information granularity. The main purpose is to fully absorb the semantic information of the ordinal relationship between objects to improve the performance of central clustering in identifying interleaved and imbalanced data. The proposed algorithm includes two main stages: first, the stage of determining the cluster center and second, the division stage. In the stage of determining the cluster center, the first cluster center is determined by using the number of conical information granularity in the data, and the remaining cluster centers are determined by defining the statistical measure of "fuzzy importance degree." In the division stage, we divide the points to be clustered into stable and active areas. The former quickly and accurately identifies and assigns the objects belonging to a cluster by measuring the fuzzy similarity between the objects to be clustered and the cluster center, and the latter assigns the objects in the active area by using the information of the points already assigned. This method describes the position and sorting relationship of objects that are granulated through ordinal relationships more accurately in the global environment, thereby gaining a more comprehensive understanding of the structural characteristics of the data. This helps to improve the accuracy and stability of clustering algorithms in handling interleaved and imbalanced data. This article uses three clustering validity indicators to test the performance of our algorithm. We compare the results with those of six different types of popular clustering algorithms and new algorithms proposed in recent years. The experimental results show that the algorithm proposed in this article can identify clusters more accurately on the datasets with a complex and staggered distribution. It is significantly better than the clustering algorithm participating in the comparison and has go

关键词： clustering algorithms Granular computing Semantics Fuzzy systems Data mining clustering methods Data models Bidirectional conical information granularity k-bidirectional conical information granularity center-based clustering fuzzy importance degree (FID) two-step division

来源：评论

学校读者我要写书评

暂无评论

Fast Characterization of Optically Detected Magnetic Resonance Spectra via Data clustering

引用

JOURNAL OF PHYSICAL CHEMISTRY C 2024年第31期128卷 13147-13154页

作者： Stone, Dylan G. Whitefield, Benjamin Kianinia, Mehran Bradac, Carlo Trent Univ Dept Phys & Astron Peterborough ON K9L 0G2 Canada Univ Technol Sydney Sch Math & Phys Sci Ultimo NSW 2007 Australia Univ Technol Sydney ARC Ctr Excellence Transformat Meta Opt Syst Ultimo NSW 2007 Australia

Optically detected magnetic resonance (ODMR) has become a well-established and powerful technique for measuring the spin state of solid-state quantum emitters, at room temperature. Relying on spin-dependent recombination processes involving the emitters' ground, excited and metastable states, ODMR is enabling spin-based quantum sensing of nanoscale electric and magnetic fields, temperature, strain and pressure, as well as imaging of individual electron and nuclear spins. Central to many of these sensing applications is the ability to reliably analyze ODMR data, as the resonance frequencies in these spectra map directly onto target physical quantities acting on the spin sensor. However, this can be onerous, as relatively long integration times & horbar;from milliseconds up to tens of seconds & horbar;are often needed to reach a signal-to-noise level suitable to determine said resonances using traditional fitting methods. Here, we present an algorithm based on data clustering that overcomes this limitation and allows determining the resonance frequencies of ODMR spectra with better accuracy (similar to 1.3x factor), higher resolution (similar to 4.7x factor) and/or overall fewer data points (similar to 5x factor) than standard approaches based on statistical inference. The proposed clustering algorithm (CA) is thus a powerful tool for many ODMR-based quantum sensing applications, especially when dealing with noisy and scarce data sets.

关键词： clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Autoencoder-Enhanced clustering: A Dimensionality Reduction Approach to Financial Time Series

引用

IEEE ACCESS 2024年 12卷 16999-17009页

作者： Cortes, Daniel Gonzalez Onieva, Enrique Lopez, Iker Pastor Trinchera, Laura Wu, Jian NEOMA Business Sch F-76825 Mont St Aignan France Univ Deusto Fac Engn Bilbao 48007 Spain

While Machine Learning significantly boosts the performance of predictive models, its efficacy varies across different data dimensions. It is essential to cluster time series data of similar characteristics, particularly in the financial sector. However, clustering financial time series data poses considerable challenges due to the market's inherent complexity and multidimensionality. To address these issues, our study introduces a novel clustering framework that leverages autoencoders for a compressed yet informative representation of financial time series. We rigorously evaluate our approach through multiple dimensionality reduction and clustering algorithms, applying it to key financial indices, including IBEX-35, CAC-40, DAX-30, S&P 500, and FTSE 100. Our findings consistently demonstrate that incorporating autoencoders significantly enhances the granularity and quality of clustering, effectively isolating distinct categories of financial time series. Our findings carry significant ramifications for the financial industry. By refining clustering methodologies, we set the stage for increasingly accurate financial predictive models, offering valuable insights for optimizing investment strategies and enhancing risk management.

关键词： Time series analysis Principal component analysis clustering algorithms Memory management clustering methods Dimensionality reduction Decoding Data compression Financial services data compression financial data processing neural network applications time series

来源：评论

学校读者我要写书评

暂无评论

Cloud-Assisted Privacy-Preserving Spectral clustering Algorithm Within a Multi-User Setting

引用

IEEE ACCESS 2024年 12卷 75965-75982页

作者： Xu, Lida Cheng, Xiangguo Tian, Weizhong Wang, Huanli Zhang, Yan Qingdao Univ Coll Comp Sci & Technol Qingdao 266071 Shandong Peoples R China Shenzhen Technol Univ Coll Big Data & Internet Shenzhen 518118 Guangdong Peoples R China Qingdao Univ Technol Sch Environm & Municipal Engn Qingdao 266520 Peoples R China Qingdao Univ Sci & Technol Coll Electromech Engn Qingdao 266061 Peoples R China

Spectral clustering, a powerful algorithm in the field of AI, holds a significant role despite its inherent high time complexity. For data owners grappling with limitations such as small datasets and restricted computational resources, harnessing the computational capabilities of cloud computing and aggregating data from multiple sources can yield precise spectral clustering results. However, explicit data uploading to cloud servers poses privacy risks. In response to this challenge, we explore the outsourcing dilemma of spectral clustering in a cloud and multi-user environment and propose a quantum-secure and efficient solution. Specifically, by employing the CKKS homomorphic encryption algorithm within a dual non-collusive server model, we formulate a comprehensive and multi-user spectral clustering outsourcing scheme. Our approach addresses privacy concerns by introducing secure computation protocols for L-2 norm, exponential function, and negative half power function. We elaborate on efficient computational algorithms for each stage of spectral clustering, ensuring accurate clustering outcomes without compromising dataset privacy. Moreover, in our scheme, users only need to upload their encrypted dataset without requiring direct interaction with each other or the cloud server until obtaining clustering results. Finally, we argue the IND-CPA security of our design and substantiate its accuracy and efficiency through theoretical comparison analysis and experimental evaluations.

关键词： clustering algorithms Outsourcing Cloud computing Servers Computational modeling Protocols Covariance matrices Homomorphic encryption Spectral analysis CKKS computation outsourcing homomorphic encryption privacy-preserving spectral clustering

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Divergence Weighted Ensemble clustering With Spectral Learning Based on Random Projections for Big Data

引用

IEEE ACCESS 2024年 12卷 20197-20208页

作者： Lahmar, Ines Zaier, Aida Yahia, Mohamed Ali, Tarig Boaullegue, Ridha Univ Gabes MACS Lab Gabes 6029 Tunisia Univ Carthage Tunis InnovCom Lab Tunis 1002 Tunisia Univ Tunis El Manar ENIT SYSCOM Lab Tunis 1002 Tunisia Amer Univ Sharjah GIS & Mapping Lab Sharjah U Arab Emirates

In many real-world applications, data are described by high-dimensional feature spaces, posing new challenges for current ensemble clustering methods. The goal is to combine sets of base clusters to enhance clustering accuracy, but this makes them susceptible to low quality. However, the reliability of present ensemble clustering in high-dimensional data still needs improvement. In this context, we propose a new fuzzy divergence-weighted ensemble clustering based on random projection and spectral learning. Firstly, random projection (RP) is used to create various dimensional data and find membership matrices via fuzzy c-means (FCM). Secondly, fuzzy partitions of random projections are ranked using entropy-based local weighting along with Kullback-Leibler (KL) divergence to detect any uncertainty. Then it used to evaluate the weight of each cluster. Finally, we create regularized graphs from these membership matrices and use spectral matrices to estimate the affinity matrices of these graphs using fuzzy KL divergence anchor graphs. Subsequently, obtaining the final clustering results is considered as an optimization problem, and the ensemble clustering results are obtained. The experimental results on high-dimensional data demonstrate the efficiency of our method compared to state-of-the-art methods.

关键词： Matrix converters Entropy clustering algorithms Uncertainty Reliability Weight measurement Sparse matrices Ensemble learning Fuzzy systems Spectral analysis Fuzzy ensemble clustering high-dimensional data random projection Kullback-Leibler divergence entropy spectral learning

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：