Malicious Android Application Package(APK) Detection Using Deep Learning

Abstract

As of late, the uses of advanced mobile phones are expanding relentlessly and   furthermore development of Android application clients are expanding. Because of development of Android application client, some gate crashers are making vindictive android application as instrument to take the delicate information and data for fraud and misrepresentation portable bank, versatile wallets. There are such a large number of malevolent applications discovery instruments and programming’s are accessible. Be that as it may, a viably and productively vindictive application recognition device expected to handle and deal with new complex pernicious applications made by interloper or programmers. This paper Utilizing Machine Learning approaches for distinguishing the malignant android application. First, dataset of past pernicious applications has to be obtained with the assistance of Help vector machine calculation and choice tree calculation make up correlation with preparing dataset. The prepared dataset can foresee the malware android applications up to 93.2 % obscure/new malware portable application

The main steps performed through this framework are sketched as follows:

1. A set of features is computed for every binary file in the training or test datasets, based on many possible ways of analyzing a malware.
2. A Deep learning system based firstly on one-sided perceptron’s, and then on feature mapped one-sided perceptron’s and a kernelized one-sided perceptron’s, combined with feature selection based on the F1 and F2 scores, is trained on a medium-size dataset consisting of clean and malware files. Cross-validation is then performed in order to choose the right values for parameters. Finally, tests are performed on another, non-related dataset. The obtained results were very encouraging.