Gini Coefficient Nonparametric Kernel Estimation Method With Application

At present,China’s economic development situation is very good,but polarization has become an urgent problem in the development process.The settlement of the gap between the rich and the poor contributes to China’s national economy,people’s livelihood,and social stability.The Gini Coefficient is an important indicator for measuring the income gap of a country or region.There are a lot of research about the Gini coefficient,and most of the studies are based on the method of parameter estimation.Besides there are many differrent methods studying it.All these method’s aim is to calculate the Gini coefficient more accurately.Non-parametric estimation is an estimation method based on sample data.The distribution of the population is directly estimated by the sample data without any assumption about its distribution.Some scholars use the estimation method to estimate the distribution function of income,and compare it with some parameters method.Because of this,the paper extends the one-dimensional nonparametric estimation to multidimensional kernel density estimation.This paper shows that the Gini coefficient is not only an index used to characterize the differences by the income distribution,but it is also an important indicator used by various countries or regions to judge the difference between rich and poor.However,the difference between rich and poor is judged not only on the one hand,but on the other such as comprehensive manifestation of the income,consumption and other information.Therefore,this paper deduces the multidimensional estimation and uses the 2014 national survey of people’s livelihood to calculate the Gini coefficient of the country,the central city,different levels of education,and different age groups.This article consists of five chapters.First of all,it introduces the research background and purpose of the study.It mainly focuses on the history of the income distribution system since the founding of the People’s Republic of China and the current status of China’s income distribution,and points out the innovations and deficiencies of this article.This chapter briefly introduces the chapter distribution and research framework.Second,it summarizes and sorts out the current research literature about the Gini coefficient.In the third chapter,the estimation function of one-dimensional kernel density estimation and multidimensional kernel density estimation is mainly deduced,and the different calculation methods of window width are discussed.Besides,how the algorithm implemented is introduced.The fourth chapter is the empirical part of this paper.The main content is the calculation of multidimensional kernel density estimation and the calculation of the conditional Gini coefficient of city,education level and age,and some simple analysis of the calculation results.In the last chapter,the full-text content and empirical analysis are summarized.At the same time,the issues that can be further studied is proposed.In general,this paper calculates the Gini coefficient under multidimensional situations and provides a new way to calculate the Gini coefficient.At the same time,it calculates the conditional Gini coefficient under one-dimension and multidimension,and can better understand the difference between rich and poor within different group.The Gini coefficient calculation derivation under continuous conditions is given in the outlook section,which provides a possible direction for subsequent research.