A Novel Semi-Supervised Learning Method Based on Fast Search and Density Peaks

Table 9

Recognition results of optical image learning set obtained after 300 iterations of the IS²DP+S²DP algorithm when the number of samples equals 21.

Target	Learning Set	Generate labeled samples		Reject	Accuracy of each type of target(%)
		Correct	Error
cats	200	143	39	18	71.50
dogs	200	154	31	15	77.00
panda	200	165	15	20	82.50
airplanes	200	—	32	168	84.00
motorbike	200	—	40	160	80.00
faces	200	—	27	173	86.50
Overall accuracy					80.00