An Improved Water Flow Optimizer for Data Clustering

• Published in: SN computer science 2024-07, Vol.5 (6), p.715, Article 715
• Main Authors: Thakral, Prateek, Kumar, Yugal
• Format: Article
• Language: English
• Subjects: Algorithms; Cluster analysis; Clustering; Clusters; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Datasets; Effectiveness; Flow mapping; Heuristic; Heuristic methods; Information Systems and Communication Service; Linear functions; Multimedia; Optimization; Original Research; Pattern Recognition and Graphics; Performance evaluation; Searching; Software Engineering/Programming and Operating Systems; Statistical tests; Vision; Water flow
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-024-03048-0

Recently, various meta-heuristic algorithms have been considered to allocate the data into different clusters based on similar information. These algorithms have obtained state of the art clustering results compared to traditional algorithms and proven their capability in the field of data clustering. This work presents an improved version of the water flow optimizer, called the IWFO algorithm for effective cluster analysis. The proposed IWFO algorithm handles the performance issues associated with the water flow optimizer algorithm such as random initialization, unbalanced search mechanism and local optima. The random initialization issues are handled through the gaussian map that can generate the initial population systematically. The search mechanism of the WFO algorithm is enhanced using the combination of non-linear functions and the previous best solution. The local optima issue is alleviated by using a neighbourhood search mechanism. The efficacy of the proposed IWFO algorithm is evaluated using benchmark clustering datasets and results are compared with popular clustering algorithms. The simulation results are assessed using intra-cluster distance (intra), standard deviation (SD), rank, accuracy rate (AR) and detection rate (DR) parameters. Some statistical tests are also performed to validate the efficiency of the proposed IWFO algorithm. The proposed IWFO algorithm improves the clustering results (average accuracy rate of more than 7%) compared to the original WFO.