Potential destination discovery for low predictability individuals based on knowledge graph

Travelers may travel to locations they have never visited, which we call potential destinations of them. Especially under a very limited observation, travelers tend show random movement patterns and usually large number destinations, make them difficult handle for mobility prediction (e.g., destination prediction). In this paper, develop new knowledge graph-based framework (PDPFKG) discovery low predictability by considering trip association relationships between We first construct graph (TKG) model the scenario entities travelers, time information) their relationships, in introduce concept private relationship complexity reduction. Then modified embedding algorithm is implemented optimize overall representation. Based on (TKGEM), possible ranking individuals’ unobserved be chosen future can obtained calculating triples’ distance. Empirically. PDPFKG tested using an anonymous vehicular dataset from 138 intersections equipped with video-based vehicle detection systems Xuancheng city, China. The results that outperforms baseline methods overall, rankings given it strong consistency travelers’ behavior choosing aggregated level. Besides, experiments indicate performance would further improved valid data introduction. Finally, provide comprehensive discussion about innovative points methodology share some findings understandings. • A generic discovery. Handle individuals focusing information Higher ranked are statistically more likely individuals. Results could introducing data.