Statistical Analysis And Application Based On Bayesian Quantile

In the past theoretical research and application,the method of regression analysis has become one of the most important statistical methods.Bassett and Koenker proposed a new regression analysis method in 1978-quantile regression.In recent years,the method of quantile regression has been widely used and developed.Quantile regression is the extension and expansion of the least squares estimation based on the conditional mean model.It is a regression method for parameter estimation by minimizing the sum of the absolute value of the weighted residuals.Using quantile regression method for analysis can give a good explanation to the data.When performing quantile regression,the distribution of the data is not mandatory,and it is also robust to outliers.When the sample size is insufficient,quantile regression cannot get an accurate forecast estimate.When the objective function is not different,it is also impossible to use quantile regression to estimate the parameters.Therefore,this article uses Bayesian quantiles to conduct theoretical research and statistical applications based on this.According to the content of this article,the article is divided into five parts.The first part is the introduction,which mainly introduces the development process and research significance of quantile regression;introduces the research results of the world in the development process of quantile and Bayesian points.The world's research results in the development of digits;introduces the main ideas and content of the thesis.The second part introduces the basic theory of quantile regression model of Bayesian method.The third part is the practical application part.This chapter uses entropy method and Bayesian quantile estimation method to analyze the comprehensive strength of Shandong Province's economy,society and ecology.Through analysis,it is found that the development of various cities in Shandong Province is unbalanced,and corresponding policies should be formulated to promote the development of various cities.The fourth part introduces the basic theory of the combination of asymmetric exponential power distribution and Bayesian quantile.And combined with the Lasso method to conduct empirical research on the problem of population impact,it is found that the main factors affecting the population are employees,GDP and other factors,and the research content is generally consistent with the actual situation.The last part is the summary part of the paper,which summarizes the basic theory and practical application of the paper.