Knowledge Graph-Based Enhanced Transformer for Metro Individual Travel Destination Prediction

Table 2

Experimental results.

		Group 1	Group 2	Group 3	Group 4	Group 5
Hit@3	LSTM	0.383	0.386	0.403	0.490	0.404
	GRU	0.159	0.210	0.128	0.169	0.128
	CNN	0.328	0.578	0.317	0.481	0.469
	FNN	0.547	0.328	0.640	0.688	0.779
	Markov	0.386	0.466	0.489	0.550	0.598
	KG-trans	0.572	0.583	0.689	0.731	0.790
Hit@5	LSTM	0.585	0.538	0.416	0.750	0.689
	GRU	0.309	0.286	0.203	0.258	0.148
	CNN	0.372	0.588	0.426	0.492	0.476
	FNN	0.602	0.464	0.758	0.762	0.794
	Markov	0.427	0.522	0.527	0.597	0.640
	KG-trans	0.611	0.760	0.767	0.809	0.810
Hit@10	LSTM	0.633	0.605	0.470	0.755	0.796
	GRU	0.618	0.433	0.316	0.456	0.160
	CNN	0.450	0.608	0.433	0.875	0.487
	FNN	0.672	0.516	0.773	0.883	0.805
	Markov	0.486	0.589	0.600	0.659	0.690
	KG-trans	0.733	0.775	0.792	0.947	0.929