The FAVAR Model Based On Lasso Method And Its Application

With the innovation of science and technology,the coming era of big data has caused global data to show explosive growth and massive aggregation.How to extract useful information from complex data sets and establish an efficiently accurate statistical model has become the most popular research problem.The factor augmented vector autoregressive model(FAVAR)has been widely used because it can effectively overcome the deficiency of insufficient information coverage in the VAR model,especially in the economic and financial fields.However,the ambiguity in the definition of K value and factors easily lead to large deviations in model estimation.As a classic coefficient compression method,the Lasso method can realize the variable selection of the model while estimating the parameters,and effectively control the model estimation and prediction errors.Therefore,the introduction of Lasso method in FAVAR model has important research significance.This paper uses the Lasso method to discuss the problem of unrecognizable factors in the FAVAR model by transforming the model selection problem into a variable selection problem,which focuses on the K value selection and factor definition of the model.Through the numerical simulation of the FAVAR model based on the Lasso method,the thesis examines the information loss caused by inappropriate K value selection to the model under different sample sizes and different iteration times.According to the simulation test results,comparative analysis was made from the three indicators of mean,standard deviation and classification accuracy.The results show that compared with the simulation test with small K value,the model estimation deviation under the large K value test is significantly reduced,and the effect is better.In the empirical part,the paper uses the FAVAR model based on the Lasso method to study the real economic problems of China’s housing prices to the monetary policy transmission mechanism.Based on the determination of the K value range,the article uses R-PCA method and Lasso method to define the model factors.Studies have shown that the Lasso method from the perspective of variable selection is more convenient and accurate to describe the meaning of factors,and can effectively help scientific researchers make correct judgments when dealing with macroeconomic regulation problems.It further confirms the practical application value of the model and shows that it is feasible to apply the Lasso method to the FAVAR model.