Dynamic Bayesian Models

Bayes forecasting model is a method based on Bayes statistics. Compared with tradition statistical method, Bayes statistics applies transcendental information, which bases on experience and historical data majorlly. This paper applies Constant-min Model, which is very normal in Bayes Model, and using discounted factorδ(0≤δ≤1), changing model DLM{1,1,V_t,W_t} to {1,1,V,δ} making the form complex-computing procedures simple. In addition, I analyzed how to deal model in different situation that whether Variance is known or unknown, and also give the formula in these situation. By proving the formula, I tell readers the formula is rational.Using a case, I prove the accuracy of the Model. Analyzing American Export amount from 1960 to 2005 and using prior information, I forecast the export amount in 2006. Meanwhile, by comparing with regression model, we can get error of two models, and known the advantage of Bayes forecast model.Finally, I introduce the meaning of Bayes forecast model's economics application.