Research And Implementation Of Incremental Bayesian Algorithm For Mobile Phone Virus Mining Engine

In recent years, mobile phones have become a necessity of life, especially the emergence and spread of smart phones, make the phone functions from a single communication extended to more areas, and people are paying more and more attention to the safety of mobile phones. On the one hand, the number of mobile phone viruses is increasing year by year, the viruses’ destructive power and influence are growing; the other hand, research in the field of mobile security is still in its infancy, research of the defense and detection to cell phone viruses (malicious codes) is of a great practical value, it is necessary to develop a simple and efficient mining engine on the mobile phone viruses excavation.This paper proposes the incremental Bayesian thinking, designs and implements the naive Bayesian module, which is a part of the mobile phone virus mining engine based on the network traffic data. In the first, this paper make a brief introduction of the mobile phone virus characteristics and attack mode, while analysis the existing mobile phone virus detection technologies and their advantages and disadvantages. And describe several Bayesian classification models and analysis their characteristics, summarize the defects of existing Bayesian classification, propose the incremental learning and incremental learning Bayesian classification algorithm, detail the improved algorithm thinking, in-depth analysis the specific learning strategies of the improved incremental learning naive Bayesian algorithm. Then, it completes data preprocessing from two aspects, the selecting and the mapping process of characteristic properties. It introduces the selection in characteristic properties of mobile phone viruses from the two points, the selecting principles and the selecting methods, and then describes the mapping methods of discrete values and continuous values. Finally, this paper summarize the entire mining engine of mobile phone viruses, describes its function structure and determine process. And on this basis, give a detailed description to the design and implementation process of Bayesian module.The Bayesian module in this paper is based on network traffic data, as the data reflects the behavior of mobile phone virus, so it can detect unknown phone viruses. And the incremental learning Bayesian classification algorithm can make up the deficiencies of spending a lot of time and energy when restart each time to learn in traditional classifiers, and reduce the adverse reactions of adding all the data when incremental learning in traditional incremental learning. This paper also uses captured network traffic data for experimental, and the experimental results show that the native Bayesian classification with incremental learning is better than the native Bayesian classifier of non-incremental learning in testing.