This codebase is an extension of the tesseract codebase.
Relevant files are in masters directory
To explore different program analysis techniques & machine learning approaches for android malware classification to determine the relative advantages and disadvantages of each, with particular focus on concept drift. By building pipelines with different approaches in program analysis, data representations, and classification models, we can compare how different techniques affect key variables such as accuracy, cost (in terms of time / resource requirements), performance decay over time (concept drift), and understandability, among others. Using this understanding we can make suggestions as to the best approach to take depending on the use case and contrast with current existing classification approaches in the literature.