HTTP Tunnel Trojan Detection Model Based on Deep Learning

Abstract

Aiming at reducing the high False Negative rate of the existing Trojan horse detection method based on behavior, this paper utilized the sequence characteristics of tunnel Trojan communication extracted from the transport layer and the bi-directional recurrent neural network in deep learning to build a HTTP tunnel Trojan detection model. The experimental result showed that the deep learning-based detection model reduced the false positive rate of normal network traffic and improved the Trojan detection rate. We also found that the deep learning-based detection model reduced the feature selecting and data cleaning process of generating samples and improved the easy-using of the HTTP tunnel Trojan detection model.