A COMPARISON OF PERFORMANCE MEASURES FOR CLASSIFICATION METHODS IN DATA MINING

Abstract: The machine learning area is being developed much as the artificial intelligence develops. It is important to evaluate the performances of the classification methods correctly since many techniques have been developed so far. Many performance measures are also developed for evaluation of the performance of the classification method. The values of the performance measures are changed under the different situations. We review several evaluation measures derived from a confusion matrix in the paper. Then, we investigate the change of the values of the performance measures under the various situations.

Keywords: Classification, Performance, Evaluation, Confusion Matrix.