Stability comparison of feature selection algorithms using normalized rank transformation

Djatna Taufik

Morimoto Yasuhiko

電気・情報関連学会中国支部連合大会講演論文集

There are needs for evaluating rank order-based similarity between different classifiers in feature selection. Feature selection maps from data set to give further understanding about the importance of ranking in decision making within feature selection algorithms. The results are ordered rankings of training and testing data. In order to compare stability within each classifier, we deploy normalized rank transformation approach to get the degree of similarity between training and testing data set. The accuracy of the selected features is then evaluated using various classifiers.

feature selection

key performance indicators

電気工学 [ 540 ]

英語

電気・情報関連学会中国支部

Copyright (c) 2007 by Authors

[NCID] AA11138730