Research And Application Of Panel Data Model Based On Information Theory

The panel data includes the time series data and cross-section data,providing a large number of data points for the model,thereby increasing the degree of freedom of the model and reducing the collinearity between model explanatory variables^[1-2].Panel data model is a hot topic in econometrics and statistics and has a wide application prospect.As one of the important problems in statistical research,model selection method can be used to solve practical problems in statistics and other disciplines.Information theory is used in statistics to optimize model variables,such as AIC information criteria,AICc information criteria c AIC information criteria,C_P criteria,all play an important role in model selection.Therefore,it is great significance to study the variable selection of panel data model based on information theory.The first three chapters of this paper introduce the panel data model and relevant methods of information theory,the statistical model chooses the commonly used index to judge the prediction effect.The fourth chapter studies the basic properties of the mixed effect model and the fixed effect model in the panel data model by combining the AIC information criteria in the model selection information theory and the AICc information criteria,and finds the corresponding expression of information criteria.By numerical simulation of Matlab,the accuracy of the real model selected by the two information criteria is compared,and the error sum of squares is selected to compare the model fitting effect.The simulation results show that:the accuracy of AICc information criteria real mixed effect model and fixed effect model is higher than that AIC information criteria of real model.The result of error square sum shows that:the model fitting effect of AIC information criterion is better than that of AIC information criterion.It can be concluded that in the selection of panel data mixed effect model and fixed effect model based on AIC and AICc two information criteria.The effect of AICc information criterion is better than that of AIC information criterion,the model selection method can be applied to practical problems.Discuss panel data random effect model variable selection based on c AIC information criterion,the fitting effect of the model is discussed through numerical simulation of Matlab.The simulation results show that:the model fitting effect of c AIC information criterion method is better than the whole model.Therefore,the method of selecting panel data model variables based on information theory can get better results.The fifth chapter mainly studies the impact of technological innovation on economic growth in China,and the analysis results show that:in this empirical analysis,the mean square error and square sum error fitting effects of individual fixed effect model variables selected based on AICc information criterion are better than the full model effect.Therefore,the model selection method based on information theory can improve the fitting effect of the model by selecting explanatory variables of the influence of technological innovation on economic growth.