Call prediction model based on smartphone users behavior

Abstract

This paper proposes a model for the prediction of the next behavior based on the smartphone call record of a user. The data of calls includes a lot of information in addition to time point and talk time. This paper systematically classifies this information and suggests a complex model to predict the next behavior of a user. The call data has a significant meaning by its nature in the frequency analysis, trend analysis, and pattern analysis, and the data is specifically classified into the 30 items and applied to the analysis. The prediction model suggested by this paper collected and arranged the 1106 data of 3 months from the users participated in the experiment, and verified using the 100 prediction data. It showed average 85.06% of accuracy, and according to the survey simultaneously conducted, the users showed the result of being satisfied with the accuracy. © 2012 Springer-Verlag.