An Improved Method of Time-Frequency Joint Analysis of Mouse Behavior for Website User Trustworthy Authentication

With the increasingly prominent problem of information security, the research on user trustworthy authentication technology becomes more and more important. Identification and authentication methods based on user’s biological behavior characteristics have attracted widespread attention due to their low cost and difficulty in imitation, which represented by mouse dynamics. This study proposed an improved method for time-frequency joint analysis of mouse behaviors for trustworthy authentication of website users. We collected the behavior data of the user’s natural mouse operation under real website environment, and analyzed the timing and spatial characteristics of the user’s mouse movements. Based on extracting the time-frequency joint distribution characteristics and spatial distribution characteristics of the temporal signals of the user’s mouse movements, we used the random forest algorithm to establish a user’s trustworthy authentication model. Mouse behavior data of five users during twenty-eight months had been used as a case study to explore the effectiveness of this method in user trustworthy authentication. The results of case analysis showed that, comparing to the original research, the method proposed in this study significantly improved the accuracy of the website user trustworthy authentication.