Multimedia System Design and Data Storage Optimization Based on Machine Learning Algorithm

Table 4

Confusion matrix of two-SVM classifier using different kernel functions.

	Formula	Linear kernel function	Transaction kernel function	Basic kernel function	Sigmoid function
True positive (TP)	—	219	227	219	210
False negative (FN)	—	13	5	13	22
False positive (FP)		8	39	6	32
True negative (TN)		188	157	190	164
Positive sample (PE)	TP + FN	232	232	232	232
Brightness sample (NE)	FP + TN	196	1%	1%	1%
Number of positive observations (POs)	TP + FP	227	266	225	242
Number of negative observations (NOs)	FN + TN	201	162	203	186
Total number of samples (TE)	PE + NE	428	428	428	428