Research And Analysis Of Examination Paper Knowledge Master Degree Based On Decision Tree

As an important index to evaluate the effect of students’learning and the ability of teachers’ teaching, examination has been attracting lots attention from students, their parents and educators. Based on different learning phrases, evaluation of education can be divided into three categories:diagnostic assessment, formative assessment and summative evaluation, in which examination plays an important role. To be specific, examination can effectively detect their knowledge in different learning phrases, and accordingly list out the teaching plan in the next step, so as to improve the effectiveness of education.As we know, the paper is compiled on the basis of the knowledge involved in related courses, by which the mastery of knowledge can be effectively detected, thus it can be seen that it is one of the main functions of the paper to detect the mastery of knowledge. In the practical application, there are two ways for teachers to analyze the students’mastery of knowledge:one is to analyze the individual paper and detect the mastery of knowledge, the other is to analyze the overall papers and have a general understanding on their mastery. The former way is suitable for individualized teaching to understand their pockets of knowledge and improve performance after filling their pockets. However, it is unable to provide guidance for the further work of teachers due to its low university in class teaching. On the contrary, the latter way is suitable for class teaching, but requires rich experience for teachers and takes a long time to analyze the mastery of knowledge, which makes the latter way difficult to achieve in real practice.The focus of this thesis lies in how to utilize each examination at utmost and help teachers analyze the relationship between the scores of student and the mastery of knowledge. By analyzing, an objective understanding on the students’mastery of knowledge on different score rung can be obtained, to crush their pockets of knowledge one by one and ultimately improve the overall scores.To solve the problem, tree algorithm, association rules algorithm, Bias algorithm and neural network algorithm are applied by many researchers to analyze the test papers, with an aim to find out the relationship between mastery of knowledge and students’ performance. However, when the paper involves a relatively few knowledge, not a single method can find out above-mentioned relationship. In view of this situation, this thesis takes the difficulty of knowledge into consideration and make detailed analysis on the knowledge involved in the paper to determine the relationship according to the degree of difficulty. The author hopes the thesis can be used to help teachers analyze students in different score rung and improve their scores.Specifically speaking, guided by the measurement of education, the thesis firstly makes an in-depth research on the current status of education measurement and the application of the test analysis application system and data mining technology into the test analysis to clarify the problems and shortcomings; secondly, the thesis makes a brief introduction on theoretical background, respectively classical test theory and decision tree; thirdly, after the combination of classical test theory and decision tree, the thesis aims at finding out the information behind the test outcome and analyzing the quality of paper. Design of knowledge mastery analysis based on decision tree model; give the detailed description of the key processes. Finally, taking the test on carbon and carbonoxide (Unit6) from Class1Grade3, Ge’ao Middle School in Yudu County of Jiangxi Province as an example, the thesis uses classical test theory to analyze the quality of the test paper. Specific index include reliability, validity, difficulty and differential degree. After analyzing the quality of the test paper, the author decides if it is significant to make further research. Then use the modle calculation the knowledge points’information gain rate, according to the obtained numerical constructing decision tree. Structural comparison of decision tree and comparing the tree and the befor tree can find in the condition of less knowledge points increase the test difficulty dimensions can effectively improve the knowledge classification.