An algorithm based on improved Bayesian inference network model for prediction protein secondary structure

This paper analyzes the Bayesian model that is used to predict secondary structure of proteins, introduces an artificial neural networks based on this model, and then give out an improved new artificial neural network model. The author´s motivation is to refer more neighboring information of amino acid residue sequences so that higher accuracy can be obtained for predicting secondary structure of the protein. We discuss data selection, network parameter determination and network performance in researching algorithm of prediction protein secondary structure. The experimental results show that the model can well cope with the problem of predicting secondary structure of proteins.