Effects of Data Sampling Interval on Accuracy of Authentication Using Wi-Fi Information Captured by Smartphone

There is a lot of researches on personal authentication methods that use behavior information. The information can be easily collected through smartphones which most people usually carry their own. The important thing to consider when collecting the information is setting the data sampling interval. The shorter the interval is, the more data we can collect, and the more we can know human behavior from the data. On the other hand, the issues of battery consumption and privacy for smartphones tracking human behavior get serious when the tracking interval is short. Therefore, the tracking interval should be long in the behavior authentication system if the change of the interval does not affect the authentication accuracy. In this paper, we examined the effects of tracking interval on behavioral authentication accuracy where the interval is set to five minutes and one hour. Wi-Fi information captured by smartphones are used and the number of the subjects are 100 randomly selected from 16,027 in this experiment. As a result, we conclude the effects on the authentication accuracy are minor.