【作者】 王双成； 毕玉江； 裴瑱；

【Author】 WANG Shuang-cheng1,2,BI Yu-jiang2,PEI Zhen3(1.School of Mathematics and Information,Shanghai Lixin University of Commerce,Shanghai 210620,China;2.Open Economic and Trade Research Center,Shanghai Lixin University of Commerce,Shanghai 201620,China;3.School of Economics & Trade,Shanghai Lixin University of Commerce,Shanghai 201620,China)

【机构】 上海立信会计学院数学与信息学院； 上海立信会计学院开放经济与贸易研究中心； 上海立信会计学院经贸学院；

【摘要】 宏观经济指标量化影响分析是宏观经济研究的一项重要内容。针对现有的分析方法侧重时序或非时序影响,不易实现二者的有机结合问题,提出了基于识别能力计算的宏观经济指标量化影响分析的动态朴素贝叶斯网络分类器方法。在分类器中,采用高斯核函数估计属性条件密度,并为高斯核函数引入形状参数,通过形状参数的优化来提高分类器的识别能力。最后,使用动态朴素贝叶斯网络分类器进行人民币汇率等因素对进出口商品的量化影响计算,并对所得到的计算结果给予经济学解释。更多还原

【Abstract】 The quantify impact analysis of macroeconomic indicators is an important research in macro-economic field.At present,timing or non-timing effects between indicators are stressed in existing methods so that the static and dynamic information can not be consistently combined.In this paper,a method of dynamic naive Bayesian net-work classifiers is developed for the quantify impact analysis between macroeconomic indicators.The analysis is re-alized by computing recognition ability of classifier.In the dynamic naive Bayesian network classifiers,the condi-tional density of attributes are estimated by using Gaussian kernel function.And shape parameter is introduced in Gaussian kernel function.The recognition ability of classifier can be by optimizing shape parameter.Calculation of the quantitative impact of RMB exchange rate and other some factors to import and export goods is done by using dynamic naive Bayesian network classifiers and results of the calculation are explained from an economic point of view.更多还原