An Efficient Kernel Learning Algorithm for Semisupervised Regression Problems

Table 1

Performance obtained using various kernel-based methods.

Method/sample Size	( = 200, = 30)	( = 400, = 100)	( = 500, = 120)
SKM	0.090 ± 0.032	0.093 ± 0.022	0.095 ± 0.018
UKM	0.085 ± 0.052	0.082 ± 0.037	0.080 ± 0.041
SEKM		0.077 ± 0.042	0.075 ± 0.036