TABLE 1.

Effect of parameters p and q in fivefold cross validation.

	q = 0.25	q = 0.5	q = 1	q = 2	q = 4	q = 8	q = 16
p = 0.25	0.9251	0.9165	0.9178	0.9246	0.9229	0.9236	0.9244
p = 0.5	0.9253	0.9236	0.9251	0.9246	0.9254	0.9235	0.9229
p = 1	0.9240	0.9250	0.9190	0.9213	0.9234	0.9234	0.9242
p = 2	0.9214	0.9204	0.9239	0.9230	0.9251	0.9181	0.9208
p = 4	0.9215	0.9222	0.9206	0.9229	0.9241	0.9239	0.9235

Bold values: LGRSH achieves the best performance while p = 0.5, q = 4.