Unemployment Prediction Of Nanjing Based On Partial Least Squares Regression And Fisher Discriminant Analysis

Unemployment is the primary problem in the economic society of the people’s livelihood, and the imperfect market economy makes the market regulation as a post regulation. To solve the unemployment problem needs remedies afterwards as well as the forecast, in order to make early preparing policies, and to avoid the economic and social shocks. Whether in theory or in practice, a perfect unemployment early-warning system should include two parts, that is unemployment forecast and unemployment alarm.This study focuses on the unemployment forecast, using partial least squares regression method and the Fisher discriminant analysis as two main tools for empirical analysis, and also two models were integrated to learn from each other to improve prediction accuracy. Firstly, establish unemployment index system, to determine the scope of variables and dependent variables. Then, in the construction of regression model, total of ten variables are selected to explore the relationship between these variables, and to predict the future unemployment. In the analysis of Fisher linear discriminant, the level of future unemployment are forecasted.Lastly, this sdudy summarizes and discusses the shortcomings of unemployment prediction, and puts forward the prospects.