Dependence Analysis And Fuzzy Forecasting Of CPI Based On Copula Function

CPI is an important economic indicator that reflects the change situations of the residents’consumption level.The magnitude of country’s macroeconomic regulation and control policy will be affected directly by the size of the CPI.PPI is a index to measure the changes of enterprise production cost as well as the important basis of economic calculation.The quantitative analysis to those two indexs is one of the hot points in today’s statistical research.In this paper,on the one hand,it studies the dependence between the CPI and PPI based on copulas function,which is beneficial to analyze the development trend of overheating or tight of the economy.On the other hand,we put forward a kind of fuzzy regression combination forecasting model and put it into CPI Short-term prediction research,which to offer reliable basis and advice for shaping economic policy and to reduce the impact brought by time lag of the policy.A class of economic data presents a significant non-linear relationship.The Pearson correlation index we usually used can not reflect non-linearity correctly and objectively.As a tool for describing dependence between variables,Copula bears remarkable advantage in the presence of non-linearity.It can describe complex dependency structure between variable flexibly,comprehensively and intensively.Considering the non-linear correlation of CPI and PPI,the paper adopted the data of country month-on-month to study the dependence between CPI and PPI and carry on the study of the relativity based on Copula which has a stronger performance ability.It turned out that Gaussian Copula has a better effect in describing rank correlation.For the fitting degree on the tail correlation and original data,Gumbel Copula is better than other Copula.Due to the limitation caused by the imprecision of input and output data of classical regression model,the fuzzy regression model solved the problem effectively with the advantage of fuzzy set theory.During the prediction,using one kind of model is difficult to reveal its change rules comprehensively.Combination prediction theory can effectively take the advantage of the useful information of single model to improve the accuracy of prediction.This paper put forward a kind of fuzzy regression combination prediction method and applied it into the prediction of CPI in China.This method can synthesize the advantages of fuzzy regression and combination prediction and can give a more truthful prediction interval value.Eventually,the results of the CPI prediction showed that fuzzy regression combination model is better than every single ones in result prediction.Furthermore,it can improve the prediction accuracy of CPI and it is fit to carry though the short-term prediction to CPI.