An Intelligent Android Malware Detection Framework using Machine Learning with Statistical Significance Analysis

Abstract

Android applications have increased at a fast rate and the openness of the Android environment has put a lot of risk in terms of malware attacks, which affected the privacy of users and the security of their android devices. The traditional signature - based malware detection methods are ineffective. This paper provides intelligent framework for malware detection in Android environment, which uses statistical significance analysis. This framework uses various classification models such as Naïve Bayes, Logistic Regression, Support Vector Machine, Decision Tree, Random Forest and deep neural network. The performance of every model is measured using accuracy, precision, recall, F1score, and ROC AUC. Experimental outcomes indicate that ensemble and deep learning techniques are more efficient than traditional classifiers, deep learning model has the high accuracy of 97%.