Bayesian Multiple Imputation In Food Corporate Credit Rating Of Application Research

In2008America’s subprime crisis evolved into a global ifnancial crisis，theworld economy suffered serious blow. At the same time, China broke out "San Lumilk powder" incident, which is China’s national dairy industry brand of a collectiveslump, to the Chinese milk industry international reputation produced fatal influence.Food corporate credit dropped to the bottom, how to thoroughly cope with the crisis，becomes the focus of corporate credit construction problems.Based on the credit rating of the initial steps-data preprocessing as the startingpoint, studies fill problems of missing data sets. The past research about corporatecredit rating on the design of the model ignored the incomplete data preprocessingresearch. In fact, the comparatively mature modeling method to the processing of thedata set has certain requirements, such as data integrity, redundancy is low, therepresentative, etc. Data loss is credit data of the commonly existing problems, to itsprocessing has become a key problem in the study of credit rating. Before we can docredit rating research process, hard to avoid can meet this or that data is lacking, sohow to deal with the problem becomes the key problem of the research of thesubsequent influence. According to the previous experience and the results of thestudy, based on Bayesian theory of multiple fill method to have missing data sets offill，and then to fill in the data were statistically analyzed, and finally the datageneration into the credit rating model，to get food corporate credit condition. Thispaper USES SAS statistical software for missing data sets fill treatment, the results ofthis paper provides five group fill value, calculated respectively29food corporatecredit score ranking situation, this paper holds that, this kind of method in a certainextent overcome the single ifll method in data fill the limitation of, very good dealwith the uncertainty of the missing data, for the subsequent research has laid a solidfoundation of data.In this paper, through the introduction of packing method, the analysis of thecurrent situation of food corporate credit, this paper expounds the data packing forrating the implications of the research, the empirical research shows the results of thestudy after fill， in comparison to the traditional methods, this method hasdemonstrative role，can better applied to food corporate credit rating, even for all thesample data missing are operable.