一种基于吉布斯采样算法的马尔可夫背景模型编码的研究

Many motif-finding programs have been developed, but no program is clearly superior in all situations. In this paper, an improved Gibbs sample algorithm is proposed to discover motif with the new method markov chain background. The improved approach can overcome the local convergence of Gibbs sample algorithm and also can accelerate the speed of the convergent course. A new method is been proposed to encode the candidate motif. This method make the modeling of the background with markov chain easily and conveniently. The markov factor is adapted to reduce the error which have a great fluence on the result of the computer. The verified biologcal data are used to test the feasibility and effectiveness of improved approach. Compared with results given by the experiment, our algorithm raise effectively the accuracy, flexibility and stability for the motif discovery.