The Empirical Analysis Of Influence Factor On Tuition Pricing For Shanghai Private Middle School

In recent around 20 years,more and more freshmen in many key high schools of Shanghai are coming from many local private middle schools,this means studying in a private middle school represents that you are getting closer to those key high schools and top universities.Therefore,despite the tuition in many private middle schools are not cheap,there are still really lots of students applying to study there.Although,there are many experts and scholars did some research on colleges/universities or kindergartens'tuition influencing factors with multivariable linear regression and BP neural network,but there are really few documents and articles specifically focus on private middle school's tuition and tuition influencing factors.Considering Shanghai is a very rich Tier 1 city which has developed very well for many years in China,based on this solid economic foundation,private middle schools developed more mature than many other cities in China,and taking data availability and systematic principles into consideration,this analysis adopt literature research method,inductive method,survey study and empirical analysis,after 3 round filtering,in the end,this research pick 92 private middle schools in Shanghai as research object,taking each semester of Y2018 tuition of Grade 6 as dependent variable;and the independent variables are 11 in total,include registered capital of each school,average students in each class,campus environment,facilities,school activities,entrance examination score,curriculum arrangement,financial appropriation for compulsory education in 2017,and each semester of Y2017 tuition of Grade 6 and etc.Based on the above variables,this research choose Random Forest and Lasso each separately build model.And then based on the common 3 features they selected to conduct Lasso regression.In the end,model result showed that,there are in total 3factors mainly influenced private middle schools tuition in Shanghai,they are:each semester of Y2017 tuition of Grade 6,facilities score and average students in each class.Assume other data keeps no change,every 1%change of each semester of Y2017 tuition of Grade 6,in average,each semester of Y2018 tuition of Grade 6 will change by 0.84%;assume other data keeps no change,every 1 score change of facilities,in average,ln of each semester of Y2018 tuition of Grade 6 will change by0.065;assume other data keeps no change,every 1 student change of average students in each class,in average,ln of each semester of Y2018 tuition of Grade 6 will change by-0.002.Comparing to the other research and studies on tuition and etc.,this analysis not only has innovation on the research object,it's the 1^st time to focus on the nine-year compulsory education period private middle school tuition pricing;but also has innovation on the variable selection,the study picked not only those macro level factors like GDP,the government financial appropriation and etc.,but also picked many micro level data specific for every school,eg.Registered capital,enrollment students and etc.;besides above 2 innovations,this research also did innovation on the research methods,this is the 1^st time conduct tuition analysis based on Random Forest and Lasso.Per model result,it's feasible to use this kind of machine learning methods to research on tuition pricing in future.A reasonable tuition amount can let more people enjoy a good quality education.I believe that this research can not only provide some reference for the private middle school tuition pricing analysis framework;but also can enrich the price research and tuition research content and samples;in addition,this research also provide a completely new perspective for the variable selection and modeling.Meanwhile,this pricing model also provides a very practical reference for the government and schools'further tuition pricing work.