A method based on improved Bayesian inference network model and hidden Markov model for prediction of protein secondary structure

This work aims at predicting the secondary structure of proteins, which is a complex nonlinear-mode classified problem. It proposes an algorithm which synchronises Bayesian network and hidden Markov model. It refers more neighbouring information of amino acid residue sequences for predicting secondary structure of the protein. Moreover it discusses data selection, network parameter determination and network performance in searching an algorithm of predicting protein secondary structure. The experimental results show feasibility and validity of the algorithm.