Prediction of enzyme subclass by using support vector machine based on improved parameters

By using of the improved parameters with increment of diversity and scoring function to express the information of sequence, a support vector machine (SVM) algorithm for predicting the enzyme subclasses of the six main functional classes is proposed. And the better results are obtained. The overall Jackknife success rates in identifying the enzyme subclasses of oxidoreductase, transferases, hydrolases, lyases, isomerases, and ligases are 94.23%, 92.94%, 90.85%, 98.43%, 99.37% and 98.96%, respectively. The results indicate that our method is helpful tool for enzyme subclasses prediction.