Machine Learning-Based Framework for Malware Analysis in Android Apps

摘要: The usage of smart phones has increased by many folds in the past decade, and in the current pandemic situation, there are apps for varied activities ranging from medicine, education, recreation, finance, and so on. Smartphone users have lot of convenience at their fingertips; however, the danger that is always lurking in the background is the entry of malware through these apps. Despite the stringent security checks done on apps, the protection against malware remains debatable. In this paper, we propose a machine learning-based framework for malware analysis in Android apps. We have applied Random Forest, Decision Tree, KNN, Linear SVM, Logistic Regression, Naive Bayes, to a custom dataset of 4,684 apps, with features extracted from permissions and opcode of the apps. The accuracy with Random Forest algorithm for our framework was 92%, whereas only permission-based method gave an …