Proposal of mathematical model of prediction of human movement on sightseeing using position information

In recent years, behavior analysis by smartphone’s GPS data is being conducted. It is becoming clear that the routine behavior of city dwellers follows a certain pattern, based on preference and life rhythm, and it is possible to predict the traffic congestion, etc., by reading this behavior pattern [1]. Prediction, based on human behavior, is also possible for non-routine behavior such as sightseeing, traveling, etc. Although they are not routine activities, there are several action patterns, based on the purpose.
We determined that it is possible to predict the movement of people, using the computational model proposed in this research. However, this model only considers the movement between cities and does not take into account the movement within the city. This model can be effectively used for predicting the movement between cities, in the case of finely divided cities such as Tokyo. However, for the Shizuoka- prefecture data used in this research, only approximate information can be obtained. Further, even if there is actual movement, it is regarded as "not moving" and is difficult to compare with actual data. To predict the movement of a person, regardless of the size of the city, we may have to change the method from "movement prediction between cities" to "movement prediction due to latitude and longitude change." Thereby, it may be possible to predict the movement of people, using detailed information.