DNA information mining based on Hidden Markov Models

Use the characteristics that different structures of the protein sequence has the different distribution of its information in the Hidden Markov Model training, classify different family of proteins sequence according to different mapping information,so as to to identify the different family of proteins. Experimental results show that the average recognition rate reach 92.8%. Recognition results show that the computing time of Hidden Markov Models is not only less than the support vector machine in multi-classification problem, but also the recognition rate is higher than support vector machine, show that the special advantages of Hidden Markov Model in dealing with multi-class DNA information mining.