Research And Application Of Third-Order Hidden Markov Prediction Method

This thesis mainly studies the prediction method of third-order hidden Markov model.Firstly,it has been widely used in the field of mathematical sequence analysis,which is a powerful and reliable basis of mathematical sequence analysis.However,because the traditional first-order hidden Markov model can not represent the dependence between longer state distances,it may ignore many useful statistical features.Therefore,this thesis proposes a prediction method of third-order hidden Markov model.Secondly,this thesis improves the existing Baum Welch algorithm and Viterbi algorithm,and deduces the third-order hidden Markov model in theory,which is the main theoretical basis of this thesis.The prediction framework proposed in this thesis is based on the third-order hidden Markov model,which integrates external quantitative information and news event information,which can effectively alleviate the problem of data sparsity.When calculating the model parameters,Baum Welch algorithm is improved,so that its parameter solution is improved from local maximum solution to global maximum solution.When solving the path with the maximum probability,the Viterbi algorithm is optimized,and the path in the original algorithm is cut to reduce the redundant state.Finally,this thesis establishes a third-order hidden Markov prediction model for stock and traffic congestion respectively,and its prediction results fully prove the feasibility and superiority of the model.At the same time,it also uses the traditional SCV model,SFM model and ARIMA model to predict respectively,the prediction accuracy of each model is measured by accuracy and other evaluation indexes.The empirical results show that compared with other models,the prediction accuracy of third-order hidden Markov is better.