Detection of Android Malware Based on Deep Forest and Feature Enhancement

Detecting Android malware in its spread or download stage is a challenging work, which can realize early detection of before it reaches user side. In this paper, we propose two-stage framework based on feature enhancement and cascade deep forest. This method detect the traffic generated encrypted transmission process malware. The first realizes binary classification benign malicious software. second multi-classification different categories To enhance data representation, convolutional neural networks used to extract features stage, principal component analysis stage. Theses extracted are spliced with payload part form fusion for task. order adapt scale samples, especially small-scale sample, cascaded forest proposed construct model. model, many layers that consist base classifiers number be automatically adjusted according samples. With combinations each layer, optima accuracy archived two stages. experimental results several datasets prove effective It also suitable unknown attacks.