Refining Sparse Cell-ID Trajectory of Public Service Vehicles by Spatiotemporal Modelling

Table 2

Classification results.

		PSV	Non-PSV	Total of true	Recall (%)
Test 1	PSV	421	35	456	92.32
	Non-PSV	41	346	387	89.41
	Total of est.	462	381	843	90.98
	Precision	91.13%	90.81%	–	–
Test 2	PSV	415	41	456	91.01
	Non-PSV	25	362	387	93.54
	Total of est.	440	403	843	92.17
	Precision	94.32%	89.83%	–	–
Test 3	PSV	404	52	456	88.60
	Non-PSV	39	348	387	89.92
	Total of est.	443	400	843	89.21
	Precision	91.20%	87.00%	–	–
Test 4	PSV	402	54	456	88.16
	Non-PSV	41	346	387	89.41
	Total of est.	443	400	843	88.73
	Precision	90.74%	86.50%	–	–
Test 5	PSV	412	44	456	90.35
	Non-PSV	45	342	387	88.37
	Total of est.	457	386	843	89.44
	Precision	90.15%	88.60%	–	–