The practice of using Machine Learning Methods in detecting Malware is growing massively. The prerequisite for implementing Machine Learning methods is the input of the dataset to it. A researcher needs to create a dataset of its own for performing Malware Detection using Machine Learning. Our dataset generation process includes Android File Collection, Decompilation, and Feature Mining Phases. We have already collected 15508 Malware Files and 4000 benign files in our Android File Collection phase and decompiled them in the Decompilation phase. Here we are discussing our Feature Mining Phase. So our goal in this paper is to select appropriate features for dataset generation. For the selection of proper features, we have also performed a Static Analysis process using online Malware Scanners. By using our static Analysis process we have selected a total of 215 features. Here we also propose the process of automating the Feature Mining from the APK files. We also have developed and implemented a Feature Mining Script in Python. Using the automated Feature Mining Script we have generated a final dataset of 16300 files. We have also discussed the working flow of feature mining script and in this paper.

APK file Malware Dataset Android Machine Learning Feature Mining and Malware Detection