一种新的高阶多元马尔可夫链模型及其应用

A New Higher-order Multivariate Markov Chain Model and Its Application

马尔可夫链广泛应用于金融、互联网、语言识别、微生物基因等领域。因此提出了一种新的高阶多元马尔可夫链模型,并给出了新模型的参数估计方法,其估计参数仅n(2s+1)+s个,少于传统模型的估计参数个数n(s+1)~2。另外,本文通过数值实验说明了新模型不仅保持了传统模型的预测性能,而且在节省计算资源方面具有较大的优越性。

Markov chains are widely used in the fields of finance,internet,speech recognition,microbial genes,and so on.In this paper,a new higher-order multivariate Markov chain model together with the associated the parameter estimation method is proposed.The new higher-order multivariate Markov chain model only needs to compute parameters,which is less than the traditional model required to evaluate parameters.Furthermore,numerical examples show that the new model not only maintains the predictive performance of traditional models,but also presents a great advantage in saving computing resources.