2025 Vol.8 Aug.N04

• Title: Peak clustering algorithm for unbalanced temporal data with deep learning support
• Name: Shiying Gao
• Company: Liaoning Institute of Science and Technology,benxi 117004 China
• Abstract:

  Peak clustering algorithms for unbalanced time-series data usually calculate the local density of each sample point in the dataset and the distance to the high-density point, and filter the clusters by manually setting a threshold and cluster the samples according to their distances to the clustering centers. Due to the lack of adaptive determination mechanism of the clustering center, it is difficult to adapt the manually set threshold to the complex characteristics of different datasets, resulting in poor clustering accuracy. In this regard, the peak clustering algorithm for unbalanced time series data under the support of deep learning is proposed. The high-dimensional complex non-equilibrium time series data are mapped to a low-dimensional feature space by using a sparse encoder to automatically learn the sparse features in the data. The decision value of a variable defined by the product of local density and distance is introduced, and the values of all sample points are arranged in descending order to generate a sequence. Mutant points are identified using a specific model based on the variation in decision values in the sequence. Pseudo-cluster centers with the characteristics of large density and small distance or small density and large distance are eliminated from the candidate cluster centers, and the objective function is constructed by calculating the affiliation of the samples to the clusters using the adaptively determined initial cluster centers and the number of clusters as inputs. The clusters to which the samples belong are determined based on the affiliation matrix to realize the data peak clustering. In the experiments, the clustering accuracy of the proposed method is verified. It is clear from the test and comparison results that the average profile coefficient of the clustering results is 0.75 when the proposed method is used for data peak clustering, which possesses a more desirable clustering effect.

  
  

• Keyword: deep learning; sparse encoder; unbalanced time series data; peak clustering;
• DOI: 10.12250/jpciams2025090805
• Citation form: Shiying Gao.Peak clustering algorithm for unbalanced temporal data with deep learning support[J]. Computer Informatization and Mechanical System,2025,Vol.8,pp.22-26

[1] Deepajothi. S ,Juliana.Survey of Clustering Algorithm of Weka Tool On Labor Dataset[J].Advances in computational sciences and technology, 2023,. 16(1):1-9.

[2] Cevikalp H , Chome E .Robust and compact maximum margin clustering for high-dimensional data[J].Neural computing & applications, 2024, 36(11) :5981-6003.

[3] Li D , Xu S , Ai X B .Data Collection in Laser-Powered UAV-Assisted IoT Networks: Phased Scheme Design Based on Improved Clustering Algorithm[J].IEEE Transactions on Green Communications and Networking, 2024, 8(1):482-497.

[4] Sinha A , Jana P K .Improved affinity propagation clustering algorithms: a PSO-based approach[J].Knowledge and Information Systems, 2025, 67(2). 1681-1711.

[5] Wang N .Design of an Intelligent Processing System for Business Data Analysis Based on Improved Clustering Algorithm[J].Procedia Computer Science , 2023, 228:1215-1224.

[6] Xiong J , Zang W , Liu Z X .Density peaks clustering algorithm with connected local density and punished relative distance[J]. supercomputing, 2024, 80(5):6140-6168.

[7] Jian W , Jingyi Z , Hua S ,et al. New Density Clustering-Based Approach for Failure Mode and Effect Analysis Considering Opinion Evolution and Bounded Confidence[J].Chinese Journal of Systems Engineering and Electronics, 2024, 35(6):1491-1506.

[8] Wang Y , Ma W J , Yang M ,et al.Unsupervised wood species identification based on multiobjective optimal clustering and feature fusion[J].Wood Science and Technology, 2025, 59(2):1-25.

[9] Priyadarshini B , Reddy D .Modified remora optimization based matching pursuit with density peak clustering for localization of epileptic seizure onset zones[J].Evolving Systems, 2023:1-17.

[10] Zhang H , Liu Y , Chao H .Density peak clustering based on improved dung beetle optimization and mahalanobis metric[J].Journal of Intelligent & Fuzzy Systems: applications in Engineering and Technology, 2023, 45(4):6179-6191.

[11] Dong Y , Dai X , Zhang Y ,et al. A Lightweight Convolutional Neural Network Based on Joint Correlation Distance Constraints and Density Peak Clustering for Hyperspectral Target Detection[J].IEEE Transactions on Geoscience and Remote Sensing, 2023, 61:1-14.

[12] Wang H , Liu B , Zhao H ,et al. Improving density peak clustering on multi-dimensional time series: rediscover and subdivide[J]. Information Systems, 2025, 67(2):1573-1596.

[13] Li M , Ma Y , Wang H B .Expanded relative density peak clustering for image segmentation[J].Pattern analysis and applications: paa, 2023, 26(4):1685 -1701.

[14] Shah S N S Z , Rosli M M .Clustering algorithms for analysing electronic medical record: a mapping study[J].IAES International Journal of Artificial Intelligence, 2023, 12(4):1784-1792.

[15] Amagata D , Hara T .Efficient Density-peaks Clustering Algorithms on Static and Dynamic Data in Euclidean Space[J].ACM transactions on knowledge discovery from data, 2024, 18(1):2.1-2.27.