Performance Evaluation of Genetic Algorithm Selection Methods in Outlier Detection: Further Analysis

Feature selection is very crucial in the activities of soft computing algorithms for quality, precision and accuracy. This paper evaluates performance some feature methods Genetic Algorithm outlier detection on fingerprint images. Roulette wheel, Rank Tournament were considered selected features enhanced using histogram equalization. K-nearest neighbor algorithm was employed classification to detect outliers. The implementation experiment carried out Matrix laboratory environment. Performance evaluated based metrics accuracy, specificity, computation time.