Development of integrated algorithm of fuzzy clustering with entropy maximization and annealing

2015 Fiscal Year Final Research Report

• Project/Area Number: 25330297
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: Gifu National College of Technology
• Principal Investigator: YASUDA Makoto 岐阜工業高等専門学校, その他部局等, 教授 (80353275)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: ファジィクラスタリング / ファジィｃ平均法 / エントロピー最大化 / アニーリング / Tsallisエントロピー / Shannonエントロピー / ファジィエントロピー

Outline of Final Research Achievements

By applying deterministic annealing to FCM maximized with Tsallis entropy, a DA-FCM algorithm has been developed. One of the challenges of this method is to determine an appropriate initial annealing temperature and a q value, according to the data distribution. Quantitative relationships between the temperature and q are examined, and it is confirmed that the temperature should be inversely proportional to q. Based on this result, a q reduction algorithm is developed in which q is defined as an inverse of a decreasing pseudo-temperature. Experiments are performed, and it was confirmed that, in many cases, appropriate q value is determined automatically from the temperature. Furthermore, the proposed methods improve the accuracy of classification and are superior to the conventional method.
Characteristics of FCM combined with entropy maximization and annealing methods are also examined, and dependencies of shapes of membership functions according to the temperature are clarified.

Free Research Field

ソフトコンピューティング