Classification Of The Difficulty Of High School Mathematics Questions Based On Random Forest

With the widespread popularity of "personalized" labels,its application to the field of education has become the focus of all walks of life.To achieve the "individualization" in the field of education,whether it is the individualization of the free test papers or the personalization of the test questions,we must first determine the difficulty of the test questions,so the research goal of this paper is to find new ways to solve the test-based questions.The difficulty of classification problems and improve the accuracy of classification.In recent years,the random forest has excellent performance in the regression and classification of other fields,and it has also brought new research ideas to the problem of “individualization” in the field of education.In this paper,the random forest algorithm is used to explore the classification problem of high school mathematics questions,and the parameters are improved and optimized based on the classical random forest algorithm.Based on the model,the high accuracy index classification difficulty algorithm is implemented.The main work of this paper is as follows:The random forest algorithm is used to classify the mathematics test questions of college entrance examination according to the degree of objective difficulty,which lays a foundation for the personalized recommendation and free test paper system.Firstly,the paper analyzes the difficulty of the test questions and the domestic and foreign research status of the classification algorithm.Secondly,it collects the mathematics test data of the college entrance examination,marks each attribute feature,and uses matplotlib package and matlab to realize the data visualization and analyze the data distribution.Calculate the correlation coefficient and consider which attributes have significant influence on the difficulty of the test,and delete the irrelevant attributes.Finally,construct the decision tree model,implement random forest based on R language,and use the training set to carry out model training,parameter tuning,and finally input the test set.The trained model is tested and the preliminary experimental results are given.The feasibility of the method is verified.The experimental results are compared with naive Bayes,k-nearest neighbor and decision tree classification algorithm.It is verified that the model can improve the high school mathematics test classification accuracy.