An novel method of protein secondary structure prediction based on compound pyramid model

In this paper, we propose a compound pyramid model to predict protein secondary structure, where homology analysis and an improved support vector machine (SVM) technology are used for predicting protein secondary structure. The homology analysis is based on BP network model which uses pair-wise sequence alignment, and SVM classification considers the physical and chemical properties of amino acids. We employed SVM multi-classification and homogenous analysis methods in integrative layer of compound pyramid model proposed by us. Result shows that the ensemble prediction model gets better results in our experiment compared with other methods.