Abstract
Bus-Net, the public transit navigation system developed by Tottori University has been introduced in Tottori prefecture. The system accumulates the log data on departure place and destination searched by users as well as designated schedule times. This study, defining users' origin-destination information stored in the path finding system as trip desire, analyzes their traffic characteristics using the trip desire. Specifically, this study clarifies the characteristics on departure place, destination, and day of the week of the trip desire by applying non-negative tensor factorization the data is from search history of Tottori city in Japan stored from October to November in 2014. As results, this study shows the difference in preferred departing time between weekdays and weekends, and characteristics such as trip desire of commuting in the center of Tottori city.
