2012 ©
             Publication
Journal Publication
Research Title Chaotic Mountain Gazelle Optimizer (CMGO): A RobustOptimization Algorithm for K-Means Clustering of Diverse Data Types 
Date of Distribution 27 November 2023 
Conference
     Title of the Conference 1st International Conference on Data Science & Artificial Intelligence 
     Organiser Asian Institute of Technology 
     Conference Place Berkeley Hotel Pratunam, Bangkok, Thailand 
     Province/State กรุงเทพมหานคร 
     Conference Date 27 November 2023 
     To 29 November 2023 
Proceeding Paper
     Volume
     Issue
     Page 18-34 
     Editors/edition/publisher  
     Abstract Addressing challenges in data clustering for diverse data types, we introduce the Chaos Mountain Gazelle Optimizer (CMGO). This enhanced Mountain Gazelle Optimizer (MGO) is tailored for K-means clustering solutions. Noticing a skew in MGO's strategy distribution, we integrated a chaotic map into the Territorial Solitary Males strategy and omitted the Migration to Search for Food strategy. This adjustment increases exploration and curtails exploitation, improving CMGO's effectiveness in clustering complex datasets. We implemented the Gower distance technique to navigate K-means clustering's limitations with categorical and binary data. Tests on numeric, binary, categorical, and mixed data underscore the clustering's versatility. We evaluated CMGO against 14 algorithms on 28 UCI and OpenML datasets using the F-Measure metric and the tied rank test for statistical significance ranking. CMGO outperforms the original MGO and other tested algorithms in clustering pure numeric and categorical data, securing first place, and third for mixed data. Thus, CMGO emerges as a robust, efficient K-means optimizing method for complex, diverse datasets. 
Author
625020038-5 Mr. TANATIP WATTHAISONG [Main Author]
College of Computing Master's Degree

Peer Review Status มีผู้ประเมินอิสระ 
Level of Conference นานาชาติ 
Type of Proceeding Full paper 
Type of Presentation Oral 
Part of thesis true 
Presentation awarding false 
Attach file
Citation 0