TABLE 3.

Training time (left) and prediction time (right) in seconds of MAPIC and competitors. Among shapelet-based classifiers, bold indicates the best performer and italic the second one (the lower the better).

Dataset	Training time						Prediction time
	MAPIC	mpdt	opt	bf	dtc	knn	MAPIC	mpdt	opt	bf	dtc	knn
ArrowHead	6.31	2.65	43.09	35.69	<0.01	0.01	0.56	5.07	0.06	6.44	<0.01	0.03
BirdChicken	3.10	2.37	56.15	34.24	<0.01	0.01	0.04	0.60	0.04	25.86	<0.01	0.02
Coffee	3.48	2.23	32.66	22.93	<0.01	0.01	0.04	0.81	0.04	7.66	<0.01	0.01
Earthquakes	1539.07	1335.27	416.79	4654.63	0.06	0.05	6.52	25.98	0.06	2.77	0.01	0.08
ECG5000	86.28	29.22	123.38	982.01	0.04	0.02	42.63	140.22	0.37	8.83	0.04	0.93
ECG200	16.11	4.70	31.08	27.15	<0.01	<0.01	0.58	2.75	0.05	1.68	<0.01	0.01
FaceFour	49.01	44.98	45.85	31.78	<0.01	0.01	0.64	11.62	0.05	11.75	<0.01	0.02
GunPoint	7.46	2.82	28.91	19.90	<0.01	0.01	0.68	4.18	0.05	3.04	<0.01	0.02
Phalanges	647.55	217.69	199.37	3552.25	0.07	0.05	15.46	43.92	0.09	1.71	0.01	0.24
PowerDemand	6.89	2.23	26.67	0.76	<0.01	<0.01	2.74	24.24	0.10	0.83	<0.01	0.05
Strawberry	122.82	43.59	236.84	3821.50	0.06	0.05	4.29	11.70	0.07	6.57	0.01	0.12
Trace	19.70	7.56	67.65	192.35	0.01	0.01	0.61	2.92	0.05	7.26	<0.01	0.02
TwoLeadECG	2.39	1.09	22.81	3.01	<0.01	0.01	1.88	30.73	0.10	1.80	0.01	0.07
Wafer	576.50	359.74	213.21	4140.29	0.06	0.07	129.88	431.18	0.49	12.91	0.06	1.59
Wine	9.85	3.88	44.13	48.80	<0.01	0.01	0.40	1.52	0.04	0.21	<0.01	0.02
WormsTwoClass	58.31	34.94	671.60	4772.95	0.05	0.06	0.90	2.51	0.07	4.69	0.01	0.07
Yoga	77.09	30.16	294.04	2855.68	0.05	0.05	19.76	94.55	0.49	59.70	0.08	0.87
Avg	109.11	124.00	150.25	1482.11	0.02	0.03	13.39	49.08	0.13	9.63	0.01	0.24
Rank	4.35	3.29	5.05	5.29	1.35	1.64	4.38	5.64	2.64	4.94	1.0	2.38