The application of data mining techniques and feature selection methods in the risk classification of Egyptian liver cancer patients using clinical and genetic data

Abstract

Data mining techniques has shown great potential in biomedical and health care fields. The objective of this paper is to apply feature selection methods and data mining techniques to Egyptian liver cancer patients' data to predict their prognosis and extract important features that affect the patient's survivability. Genetic and Clinical data from 1541 patients were analyzed. Three feature selection methods and seven data mining techniques were studied and compared. Wrapper Subset method and Random Forest proved to be the best performing feature selection method and data mining technique respectively. Moreover, important genetic features such as p53 gene exon 6 and 9 mutations proved to have a significant impact on patient's overall prognosis.