经济增长影响因素分析的动态贝叶斯网络方法

The method of dynamic Bayesian networks for economic growth impacting analysis

给出了影响经济增长相关因素分析的多马尔科夫链动态朴素贝叶斯分类器方法,该方法基于属性对类的识别能力进行指标之间的影响计算.在分类器中,属性和类均构成马尔科夫链,并通过朴素贝叶斯网络结构将这些马尔科夫链组合在一起形成分类器结构,这将使相关指标的动态和静态信息均能得到充分的利用,从而使影响分析更加可靠和实用.

The impact factor analysis of economic growth is an important issue for studying economic operation. There have been many researches on the impact factor analysis of economic growth based on quantitative economics methods. But these methods mainly use timing or non-timing information so that two kinds of information can not be integrated organically. In this paper, a method of multiple Markov chain dynamic naive Bayesian classifiers is developed for the impact factor analysis of economic growth. In the method, the recognition capabilities of attributes to class is used to affect calculation between macroeconomic indicators. And attributes and class separately form Markov chains. These Markov chains are combined with naive Bayesian network to form the classifier structure. As a result, dynamic and static information can be fully used to improve the reliability and the availability of impact analysis between macroeconomic indicators.