Feature Selection for Colon Cancer Detection Using K-Means Clustering and Modified Harmony Search Algorithm

This paper proposes a feature selection method that is effective in distinguishing colorectal cancer patients from normal individuals using K-means clustering and the modified harmony search algorithm. As genetic cause of originates mutations genes, it important to classify presence or absence through gene information. The proposed methodology consists four steps. First, original data are Z-normalized by preprocessing. Candidate genes then selected Fisher score. Next, one representative each cluster after candidate clustered clustering. Finally, carried out combination created applied classification model verified 5-fold cross-validation. obtained accuracy up 94.36%. Furthermore, on comparing with other methods, we prove performs well classifying cancer. Moreover, believe can be not only but also gene-related diseases.