A study on estimation of human personality from location visiting preference

It is widely believed that human personality and human-preferred locations are closely related. In this paper, a statistical study to determine the relationship between personality and preferred location, quantitatively, was conducted. Human personality can be represented through the big five factor (BFF) model. Human-preferred locations have been determined from a positioning dataset collected by the smartphones and other mobile devices of volunteers. The positioning dataset can be analyzed and clustered into preferred location datasets. Such analysis is composed of two phases. The first phase estimates a human personality from the dataset of human-preferred locations. The second phase shows the effect of human personality on the selection of preferred locations. In each phase, the location and personality data are preprocessed, trained using a back propagation network (BPN), and then validated and studied through a regression analysis. Location clustering, BPN training, and a regression analysis are the major methods used. The results of this paper demonstrate the relationship between personality and preferred location, which can be presented quantitatively. In particular, the estimation of human personality based on favored locations is the key to our research. For example, preference for a home location is highly related with extraversion, and rarely related with conscientiousness.