The Application Of Data Mining Technology In The Analysis Of Labor Statistics

As a long-term and high-speed development of China,the labor economy achieved continous and steady growth.Regarding on the imperfect market economic system of China,the labor economy is also facing critical challenge.There are lots of huge gaps in different regions,industries.Labor wages,working environment,welfare and other relative facts did not get enough balanced development.Therefore,it is necessary to get benefit of data mining technology to analyze the status of labor economy and predict future development.The data in this paper comes from "China Labor Statistics Yearbook 2016",which is an annual statistical publication that comprehensively reflects the labor economy in China.The data of labor statistics for all provinces,autonomous regions,and municipalities directly under the Central Government in 2015 was compiled,and the key indicators were also compiled with statistical data over the years.By studying the labor statistics in "China Labor Statistics Yearbook 2016",we will focus on labor statistics and forecasting average wages,by applying following 4 models.1?Analysis of the labor statistics indicators(Model 1,Model 2).Analysis of 14 labor statistics indicators for 31 provincial-level administrative units(excluding Taiwan,Hong Kong,and Macao)across the country.The first model is the principal component analysis model.The principal component analysis method is used to analyze the labor statistics indicators,and the principal component scores of the relevant labor statistics indicators are calculated to summarize the status of the labor economy in various regions of the country.The second model is a cluster analysis model.The classic K-means clustering method in cluster analysis was used to study the labor statistics index,and the provincial administrative units were clustered to obtain the clustering results.The status of the labor economy in different regions in China was analyzed as well.The results of the study: Model 1 or 2 can be divided into provincial administrative regions of China,and the ranking order obtained by model 1 is consistent with the result of model 2 classification.2?Predict average salary(Model 3,Model 4).Taking the wage data from 1985 to 2010 as the research object,predict the average wage of urban employment in China from 2011 to 2015,compare the results with actual values,and verify the fitting effect of the model.The third model is a multiple linear regression model.A multivariate linear regression model was established based on the variables of the wage level as a variable to predict the average wage.The fourth model is a time series analysis model.The average salary is predicted by using the time series method based on the ARIMA model.The results of the study: Model 3 or 4 can well predict average wages.By comparing the predicted value with the actual value for five consecutive years,the error distribution is within a reasonable range.