부산대학교 도서관

In this paper, a penalty function is added in the construction of the VAR model to control the accuracy of the model prediction, and considering the nature of the correlation between the variables, a weight fusion with Laplace penalty is introduced to generate a sparse solution to complete the construction of the vector autoregressive model. The basic types of vector autoregressive models under panel data are divided, the number of parameters to be estimated is determined, the parameters of the vector autoregressive model are estimated by the generalized method of moments estimation, and the parameter fitting effect of the model is improved by the great likelihood parameter estimation method, and the vector autoregressive model for panel data is constructed. Based on the VAR model, we propose a design scheme for the teaching of Civics in college sports courses and analyze the development of sports + Civics and the teaching effect of the courses through simulation and empirical analysis. The results of the study showed that OSI and OPCI rejected the original hypothesis at the significance level of 5% after second-order differencing, at which time both p-values were less than 0.1, and the series was smooth. The elasticity coefficient of the starting value of the sports response path is 0.0342, and the error interval is [0.0564, 0.06], indicating that the future development of the scale of “Sports + Civics” is stable.