Determine Fuzzy Measures in Multiple Classifiers Fusion Model

Abstract

In finite set, Choquet fuzzy integral with respect to fuzzy measures can be transferred into linear combination of product, based on this fact we can choose standard optimization technical to determine fuzzy measures. This paper present linear programming and quadratic programming to determine fuzzy measures, the experiments demonstrate that classification accuracy of fuzzy integral with respect to fuzzy measure is better than the classification accuracies of majority voting and weighted average.

Choquet integral based samples reduction in multiple classifiers combination

“…Choquet integral value is always comprised between min and max of the function values. So we can conclude the following two propositions [15] when Choquet integral is used to combine multiple classifiers in classification. A more exhaustive proof will be found in [15].…”

“…In our experiments, the reduction samples of training set refer to reduction the outputs of every classifier in training set not samples of training set directly, and the rule is similar to that in paper [15]. In Comparison the results in TABLE Ⅱ and TABLE Ⅲ and those in TABLE Ⅵ and TABLE Ⅶ, the testing accuracies of combination system RECS1 better than those of CS1 in cmc and wine datasets, the testing accuracies of combination system RECS2 are higher than those of CS2 in four datasets.…”