Gas Ito Algorithm in Applications of Data Clustering

• Main Authors: I-Ting Kuo, 郭宜婷
• Other Authors: 邱垂昱
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/up2mja

博士 === 國立臺北科技大學 === 工業工程與管理系博士班 === 102 === Recently, the progress of information technology has transformed the way of marketing and information management in companies. Data mining has been successfully applied in many fields to find useful information existed in vast database. Clustering tool is the most important applications of data mining which tries to segment data into homogeneous clusters and is one of the most useful technologies in data mining methods. Cluster analysis is a common technique for statistical data analysis used in many fields, including machine learning, pattern recognition information retrieval and bioinformatics. Clustering can be formulated as a multi-objective optimization problem. In this study, a new algorithm is proposed for clustering problems. Based on gases brownian motion optimization algorithm (GBMO), this research proposes a gases Ito algorithm (GIA) that introduces particle drift and particle diffusion model which aimed at enhancing global search capability. The proposed algorithm is evaluated based on 5 benchmark functions and 3 data sets of UCI Repository. The results are compared with recent algorithms in the literature. Comparison results with benchmark test illustrate that the proposed algorithm possesses better ability to find the global optimum than other algorithms and is an effective global optimization tool.