Research on active learning based computer viruses detection approaches

As traditional computer viruses detection approaches update slowly and have poor ability in detecting unknown viruses, active learning is well-suited to many problems in viruses detect processing, where unlabeled data may be abundant but annotationis slow and expensive. This paper aim to shed light on the application of the active learning theory in computer viruses detection. Moreover, to improve the precision of the virus detection and the efficiency of the active learning process, query function based on the uncertainty based sampling is realized. Experiments' results show that the model has very good detection precision against unknown computer viruses and can greatly shorten the training time and reduce the requirements of the training data and improve the learning efficiency of the system.