Method for Feature Selection of Machine Learning Based Malware Detection RECORDING MEDIUM and Apparatus FOR PERFORMING THE METHOD

摘要

Characteristic information of each application in Android environment ( ) Create an initial feature set (IFS), ) Are ranked in order of high relevance to malware, and the characteristic information ( The number of malicious applications that contain ) And the number of normal applications ( ), The distribution rate of the malicious application and the distribution rate of the normal application are calculated, and compared with the preset distribution rate thresholds, respectively, Rank value of ) And weights (w), and the characteristic information ( ) Is a feature selection method for machine learning based malware detection that classifies) into the final feature set (DFS).