Least Absolute Deviation Support Vector Regression

Table 2

Experiment results on Benchmark datasets.

Number	Dataset	Algorithm	RMSE	MAE	SSE/SST	SSR/SST	Time (s)	Iter	#train	#test
1	Bodyfat	LS-SVR	0.0075	0.0056	0.1868	0.7799	0.0854	/	200	52
	(252 × 14)	LAD-SVR	0.0026	0.0012	0.0277	0.9686	0.0988	4.4	200	52
2	Pyrim	LS-SVR	0.1056	0.0649	0.5925	0.5729	0.0137	/	50	24
	(74 × 27)	LAD-SVR	0.1047	0.0641	0.5832	0.5342	0.0385	9.8	50	24
3	Pollution	LS-SVR	39.8592	31.0642	0.5358	0.8910	0.0083	/	40	20
	(60 × 16)	LAD-SVR	36.1070	28.2570	0.4379	0.7312	0.0202	5	40	20
4	Triazines	LS-SVR	0.1481	0.1126	0.9295	0.3356	0.0759	/	150	36
	(186 × 60)	LAD-SVR	0.1415	0.1066	0.8420	0.3026	0.0958	6	150	36
5	MCPU	LS-SVR	53.6345	27.5341	0.1567	0.9485	0.0509	/	150	59
	(209 × 6)	LAD-SVR	50.2263	26.5241	0.1364	0.9666	0.0883	5.4	150	59
6	AutoMPG	LS-SVR	2.8980	2.1601	0.1308	0.8421	0.2916	/	300	92
	(392 × 7)	LAD-SVR	2.6630	1.9208	0.1102	0.8502	0.3815	6.4	300	92
7	BH	LS-SVR	4.3860	3.2163	0.2402	0.8563	0.2993	/	300	206
	(506 × 13)	LAD-SVR	3.6304	2.5349	0.1648	0.8593	0.4971	8.8	300	206
8	Servo	LS-SVR	0.7054	0.4194	0.2126	0.8353	0.0342	/	100	67
	(167 × 4)	LAD-SVR	0.6830	0.3767	0.2010	0.8147	0.0561	4.9	100	67
9	Concrete	LS-SVR	7.0293	5.1977	0.1740	0.9005	1.1719	/	500	530
	(1030 × 8)	LAD-SVR	6.8708	5.0748	0.1662	0.8936	2.0737	8.7	500	530