Panel Data Parameter Estimation Based On MLS

With the rapid development of information technology industry,the form of data storage is more and more convenient.Compared with the past decades,we are more likely to obtain individual long-term activity records.Compared with other traditional models,panel data model has the advantages of collecting more information and mo re obvious reflecting the differences of unobservable variables.Therefore,it appears in real life and should More and more scenes are used.However,the parametric panel data model drives a large number of researchers to study the nonparametric panel dat a model because of the strong assumption.This thesis firstly reviews a lot of previous research work on panel data model parameter estimation,analyzes the hot issues of current research,and puts forward the main research points of this thesis;secondly,it analyzes the current principle of panel data model parameter estimation and non parameter estimation,and mobile least square method has more advantages than traditional least square method Based on the advantages of MLS,a new method of nonparametric parameter estimation of panel data model is proposed,and its error estimation is obtained.In this thesis,root mean square error and mean absolute error are used as the evaluation criteria of the model,through many simulation experiments and many times in Guangdong,Jiangsu,Shandong and Zhejiang provinces The provincial panel data model composed of fiscal revenue and real estate development investment of Sichuan,Henan,Hubei,Hunan,Fujian and Shanghai from 1994 to 2016 is used to verify the reliabilit y of the new method.The results of root mean square error and mean absolute error show the feasibility of the improved method.The new estimation method has a certain value in small sample estimation of panel data.