A Comparative Study Of The Class Model In The Evaluation Of Teaching Quality Appreciation

For a long time,China has been to measure the teaching quality of schools by students' test scores,especially the high school entrance examination enrollment rate as the only standard to measure the quality of school teaching,which is an unscientific evaluation method.Value-added evaluation is a developmental evaluation based on the progress of students,which can objectively and fairly evaluate the teaching quality of the schools.Hierarchical linear model is widely used in the value-added evaluation of teaching quality due to its unique advantages in processing hierarchical data,but it can only measure the linear relationship between data and has some limitations.Therefore,this study extends the hierarchical linear model to the non-linear model,conducts a comparative study of different hierarchical models,and compares different hierarchical models to establish an appropriate value-added evaluation model to objectively measure the progress of students,and then scientifically evaluate the teaching quality of school.This paper is divided into five chapters,including the introduction of the theory,the theoretical expansion of the hierarchical model,the determination and optimization of the hierarchical model,the empirical comparison of the hierarchical model,the conclusion and thinking.In the first chapter,combined with the development status of education evaluation in China,the problems to be solved in this paper are proposed,and the purpose and significance of this paper are determined.The research on education value-added evaluation model at home and abroad is summarized,and the research content of this paper is described in combination with the technical roadmap.In the second chapter,the hierarchy model is theoretically expanded,and the general form of hierarchy model is established.On the basis of the hierarchical linear model,the parameter estimation and parameter test of the model are theoretically deduced by EM algorithm.In the third chapter,five common regression models are combined with the hierarchical model to construct a hierarchical linear model,a hierarchical power regression model,a hierarchical exponential regression model,a hierarchical inverse proportional regression model,a hierarchical quadratic regression model,and the optimal model judgment criterion of "double difference method" is studied.Chapter four conducts an empirical comparative study of the extended hierarchical model.After collecting the total score data of middle school and college entrance examination of 160 schools in a certain province,on the basis of standardization of the data,five hierarchical models were used to model and solve the data,and the optimal model under the current data was selected as the hierarchical quadratic regression model by comparison.Compared with the traditional hierarchical linear regression model,the error rate of the hierarchical quadratic regression model is reduced by 1.17%,the sum of squared residuals is reduced by 2.03%,and the educational value-added ranking of 160 high schools is calculated more scientifically and accurately.In the fifth chapter,the research on hierarchical model is summarized,and the innovation of this paper in three aspects,namely,the expansion of hierarchical model,the idea of model comparison and the comparison method of "double difference method",is illustrated.Finally,the future research work is prospected.