Multi-objective evolutionary algorithms for feature selection : application in bankruptcy prediction

摘要

A Multi-Objective Evolutionary Algorithm (MOEA) wasadapted in order to deal with problems of feature selection in datamining.The aim is to maximize the accuracy of the classifier and/or tominimize the errors produced while minimizing the number of featuresnecessary. A Support Vector Machines (SVM) classifier was adopted.Simultaneously, the parameters required by the classifier were also optimized.The validity of the methodology proposed was tested in theproblem of bankruptcy prediction using a database containing financialstatements of 1200 medium sized private French companies. The resultsproduced shown that MOEA is an efficient feature selection approachand the best results were obtained when the accuracy, the errors and theclassifiers parameters are optimized.