Article
Computer Science, Artificial Intelligence
Zhen Jiang, Yongzhao Zhan, Qirong Mao, Yang Du
Summary: Semi-supervised clustering aims to utilize prior knowledge to improve clustering performance. Existing methods do not adequately consider the natural gap between class information and clustering when using partial labeling information. In order to address this issue, a compact-cluster assumption is proposed along with a general framework called CSSC, which supervises traditional clustering using an objective function that measures the compactness of clusters. Two effective solutions for Kmeans and spectral clustering are provided within this framework. The proposed method is shown to be feasible and effective through theoretical analyses and extensive experiments on real-world datasets.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Review
Computer Science, Information Systems
Jianghui Cai, Jing Hao, Haifeng Yang, Xujun Zhao, Yuqing Yang
Summary: Semi-supervised clustering (SSC) is a technique that integrates semi-supervised learning and clustering analysis to improve clustering performance by incorporating prior information. This paper provides a comprehensive review of SSC, organized into different categories and discusses their performance, suitable scenarios, and ways to add supervising information. It also summarizes successful applications of SSC in various fields and provides application caveats and development trends. This review and analysis of SSC can benefit researchers in providing an overall understanding, research topics, and analysis of existing methods.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Bruno Vicente Alves de Lima, Adriao Duarte Doria Neto, Lucia Emilia Soares Silva, Vinicius Ponte Machado
Summary: This paper introduces a framework for data self-labeling based on deep autoencoder combined with a self-labeled technique that takes into consideration cross-entropy. Clustering learning in the reduced dimensionality space Z helps adjust the weights of the labeling layer, and the proposed method achieves competitive performance compared to classic methods found in the literature.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Computer Science, Information Systems
Si-chao Lei, Xing Tian, Wing W. Y. Ng, Yue-Jiao Gong
Summary: Image hashing methods have been proven effective and efficient for large-scale image retrieval. However, existing methods often manually select fixed hash code lengths, which may not optimize retrieval performance. In this paper, a length adaptive hashing method is proposed to optimize hash code lengths adaptively using a multiobjective evolutionary algorithm. Experimental results show significant improvement in retrieval performance compared to traditional methods.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Software Engineering
Adan Jose-Garcia, Wilfrido Gomez-Flores
Summary: We introduce CVIK, a MATLAB-based toolbox that helps with cluster analysis applications. This toolbox provides 28 cluster validity indices (CVIs) to measure clustering quality for data scientists, researchers, and practitioners. CVIK supports two approaches for automatic clustering: evaluating candidate clustering solutions from classical algorithms and assessing potential solutions in evolutionary clustering algorithms using optimization methods. It also includes different proximity measures for estimating data similarity, and can handle both feature data and relational data. The source code and examples are available in this GitHub repository: https://github.com/adanjoga/cvik-toolbox.
Article
Computer Science, Information Systems
Fuhao Gao, Weifeng Gao, Lingling Huang, Jin Xie, Maoguo Gong
Summary: This paper proposes an evolutionary multitasking optimization algorithm that transfers effective knowledge through semi-supervised learning. By utilizing labeled and unlabeled samples generated in the optimization process, the algorithm identifies individuals with valuable knowledge and transfers the knowledge between tasks, leading to significantly improved performance.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Behnam Tavakkol, Jeongsub Choi, Myong Kee Jeong, L. Susan Albin
Summary: Clustering validity indices are used to determine the correct number of clusters and evaluate the quality of clusters formed by clustering algorithms. Internal validity indices, such as OCVD, focus on capturing the separation and compactness of clusters by considering the density of data objects. OCVD, a single number that averages the density-based contribution of individual data objects, performs well in detecting the correct number of clusters, particularly in data sets with clusters of arbitrary shapes.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Tianshu Yang, Nicolas Pasquier, Frederic Precioso
Summary: A novel semi-supervised consensus clustering algorithm is proposed in this article, which utilizes closed pattern mining technique to generate a recommended consensus solution without the need for inputting the number of generated clusters k, and can improve the quality of clustering results.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Ding Zhang, Youlong Yang, Haiquan Qiu
Summary: In this paper, a two-stage semi-supervised clustering ensemble framework is proposed to address the limitations of existing semi-supervised clustering ensemble methods. The framework considers both the selection of ensemble members and the weighting of clusters to achieve a final partition with improved performance.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Zahra Ghasemi, Hadi Akbarzadeh Khorshidi, Uwe Aickelin
Summary: This study focuses on clustering problems and aims to predict outcome variables by partitioning data points into similar clusters using a multi-objective optimization approach. Local regression is used to predict the outcome variable, and the performance of the multi-objective models is compared to single-objective models.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Yongxuan Lai, Songyao He, Zhijie Lin, Fan Yang, Qifeng Zhou, Xiaofang Zhou
Summary: This article proposes a new framework that generates base partitions in an unsupervised manner and assigns different weights to each cluster of the base partitions. The weighted co-association matrix based consensus approach is then applied to achieve a final partition. Empirical results show that the new framework retains high accuracy, adaptability, and robustness.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Zhen Jiang, Lingyun Zhao, Yu Lu, Yongzhao Zhan, Qirong Mao
Summary: This paper proposes a semi-supervised hybrid resampling (SSHR) method, which captures the data distribution for both over-sampling and under-sampling by running semi-supervised clustering. The method achieves the best performances in terms of both F-measure and AUC on 44 benchmark datasets.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Hao Li, Yongli Wang, Yanchao Li, Gang Xiao, Peng Hu, Ruxin Zhao, Bo Li
Summary: Batch mode active learning (BMAL) aims to train reliable learning models by efficiently requesting ground truth annotations for beneficial unlabeled points. However, current BMAL methods may have suboptimal batch acquisition due to fixed weights for sampling criteria. This work proposes an Adaptive Criteria Weights batch selection algorithm (ACW) to dynamically adjust the importance of criteria for semi-supervised learning, demonstrating superiority over existing BMAL approaches.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Xiao Xu, Haiwei Hou, Shifei Ding
Summary: This paper proposes a novel semi-supervised deep density clustering (SDDC) method, which uses a convolutional autoencoder to learn embedded features, designs a semi-supervised density peaks clustering to identify stable cluster centers, and introduces prior information to guide the clustering process.
APPLIED SOFT COMPUTING
(2023)
Article
Remote Sensing
Haifeng Song, Weiwei Yang
Summary: A general semi-supervised scene classification method based on clustering and transfer learning is proposed to improve scene classification performance by iteratively optimizing the model.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)