Research Of Association Analysis Based On Educational Data

With the continuous development of information technology,the traditional educational research methods also need to keep pace with the times.Studying and interpreting various phenomenas and problems that exist in education by advanced technical measures is an indispensable part of modern educational research.Due to the bringing forward of big data concepts,data has begun to occupy the dominant position in the development of all walks of life.All kinds of education data analysis are becoming the powerful support tool for the views and decision that proposed by educators now.Under this background,we begin to study the curriculum relevance for teaching quality evaluation and test items association for personalized learning,we also investigate the correlation of students' achievement in learning analysis based on regression modelThe evaluation of teachers' curriculum quality mainly relies on the students' subjective consciousness at present,and there are few researches on this direction with objective view.This paper first starts mining the data of students' test scores based on Apriori algorithm and finding the relative course according to the mining result.Then we see the performance of students in relevant courses from their learning situation at the same course.We will use students' test score in relevant courses to judge the teaching quality of the courses' teacher.Personalized learning has always been the focus of educators in modern education,and its biggest advantage is to help each student get the corresponding learning guide to themselves.Because the final purpose of personalized learning is improving students' performance on course,we need to provide them suitable test items.How to choose the personlized test items need to kowning the parameters of test items well.Now the research of test items is mainly focused on the difficulty and discrimination,but there are few research studies are specifically concerned with test item associations.In Generally,most of existed research on test item associations are conducted under the premise of knowing the properties of test items.This paper proposes a test items association model which not relay on the detail of test items,we could quantify the different test items associations through the model.The impact of the model on different scenarios is revealed under a number of experiments,also the effectiveness of the model is validated based on the actual students' response record data.Based on this model,this paper designs a method of intelligent test construction,they can provide students with weak knowledge points of special exercises and breadth inspection of all knowledge points in course,which could meet the needs of students' personalized learning.With the popularity of personalized learning,learning analysis technology has begun to be widely used.This paper analyzes the relationship between students' course achievements,students' learning habits and their daily performance based on regression model.We also use the regression model to predict students' final achievements according to previous student data.At the end of this paper,we introduce the personalized learning platform that we developped,and make a brief description of the various functional modules in the platform.The platform has accumulated a large number of student data,we combine these data with the test items association model which mainly to help students get a better way directs to their own diagnostic learning.