ショケ積分モデルの構築とその応用(一般講演,ライフサイエンスとソフトンピューティング)

抄録

This paper describes a method of determining fuzzy measures for building the Choquet integral model and a medical application. Since fuzzy measures do not require additivity and their property is monotonicity, these measures can be interpreted as subjectivity measures of persons who evaluate objects. λ-fuzzy measure is utilized in the model and the densities of fuzzy measures and the parameter λ, are obtained by genetic algorithms often used in combinatorial optimization problems. An endoscopic image diagnostic model constructed by the method which shows its applicability is illustrated.