Augmented Intention Model for Next-Location Prediction from Graphical Trajectory Context

Table 2

Accuracy of top k results with various prediction methods.

Dataset	Urban travel				Foursquare NY				Foursquare TKY
Methods	acc@1	acc@5	acc@10	acc@20	acc@1	acc@5	acc@10	acc@20	acc@1	acc@5	acc@10	acc@20
Statistics	22.69	36.71	39.42	41.20	9.22	17.91	21.43	25.22	12.06	21.46	24.72	28.62
Markov chain	15.42	36.52	44.32	51.48	9.78	22.24	27.08	29.10	11.76	23.68	29.61	34.53
DeepMove (max, dot)	24.94	43.79	52.06	59.60	9.35	23.62	29.76	34.57	10.30	24.84	32.21	38.76
DeepMove (mean, dot)	23.73	42.55	51.13	56.71	9.87	23.03	28.86	31.32	9.37	22.66	30.18	34.35
DeepMove (max, concat)	24.09	43.17	51.76	59.88	10.16	21.98	26.99	30.96	10.44	24.55	31.53	36.74
AI-RNN (RS)	27.13	51.80	62.43	71.18	12.04	23.78	27.78	32.08	11.36	23.36	28.46	33.02
AI-RNN (DO)	27.65	52.42	62.87	71.95	10.23	22.19	26.45	31.72	11.07	22.87	27.08	32.13
AI-RNN (MP)	27.14	51.89	62.19	70.90	12.67	23.98	28.02	35.36	12.24	25.56	32.15	35.89