Research And Implementation Of Teaching Evaluation Model Based On Machine Learning

Teaching evaluation is a process of studying teachers' teaching and students'learning value.It has become an important part of teaching management and teaching process in colleges and universities.There are many common teaching evaluation systems,most of which evaluate teachers' behavior,while students'learning process and effects are rarely mentioned.At the same time,the implementation of teaching evaluation workflow is cumbersome,and it often need to complete a lot of data computing tasks.Therefore,how to use modern science and technology to establish a perfect,objective and feasible classroom teaching evaluation system and optimize the evaluation process is an important problem to be solved.This dissertation starts from the construction of learning-centered teaching evaluation system in colleges and universities,by using data correlation analysis,association rules and other methods to optimize the indicators of teaching evaluation by students.At the same time,the machine learning algorithm is introduced into the teaching evaluation process,and the teaching evaluation process is automated by constructing the teaching evaluation model.Firstly,for the acquired teaching evaluation data set,this dissertation adopted linear regression method and density-based outlier detection method to clean the abnormal data.In order to verify the independence of evaluation attributes,this dissertation proposed a correlation analysis method based on association rules algorithm,which judged the dependencies between attributes according to the confidence of the rules,and then determined the related items with strong correlation by combining the correlation coefficients between attributes,removing redundant items and optimizing the evaluation system.Secondly,this dissertation proposed a teaching evaluation model based on weighted naive bayes algorithm.According to the influence degree of different attributes on the evaluation results,a method for determining the weight of each evaluation attribute by using the correlation probability of the class attribute was proposed,and corresponding weights were set for each evaluation index,so as to construct a classification model suitable for teaching evaluation.Finally,in order to improve the algorithm efficiency,the weighted bayesian incremental learning method was adopted to solve the problem of batch arrival of new samples.In this strategy,it is not necessary to retrain the old sample data,but only to adjust the model parameters according to the new sample data.In order to verify the feasibility of the algorithm,this dissertation collected the teaching evaluation data of students in a university for experiment.The experimental results show that the classification accuracy of the model constructed by the weighted naive bayes algorithm can reach 75%,which is about 3%higher than the traditional bayes classification algorithm.At the same time,in the face of rapid data growth,the incremental learning method can not only improve the classification performance of the classification model,but also improve the time efficiency of the algorithm.The construction process of the teaching evaluation model has a strong theoretical and practical reference value for the research of educational informatization,and can also improve the efficiency of practical teaching evaluation.