Prediction of Antimicrobial Peptides Based on Sequence Alignment and Support Vector Machine-Pairwise Algorithm Utilizing LZ-Complexity

Table 5

The performance comparison of sequence alignment, SVM-LZ complexity pairwise, and our integrated algorithm across both training data on independent test using Wang test set.

Type of training data	Algorithm	Number of predicted sequences	Number of correctly; predicted sequences
Normal training set	Sequence alignment	1025	933	91.02%
	SVM-pairwise	All (1136)	870	76.85%
	Sequence alignment + SVM-pairwise	All (1136)	995	87.59%
<0.7 training set	Sequence alignment	1004	812	80.88%
	SVM-pairwise	All (1136)	926	81.51%
	Sequence alignment + SVM-pairwise	All (1136)	894	78.70%