The Research And Development Of Data Mining System Based On Error Test Of Middle School Students

The rapid development of internet technology promotes and accelerates the appearance of the new educational pattern. At present, a lot of network learning platform have been widely used, therefore, to manage wrong themes produced in network learning platform by students, error book of network also begins to rise and popular. Based on the data mining theory, the article researched and implemented a data mining system of error themes for middle school students. On one side,the system can help students quickly and efficiently build personalized error themes book, and classify error themes, and analysis the error of error theme book,On the other hand, through the data mining technology, the system can analysis weak points of student and recommend a set of personalized examination questions for each student,thus it reduced the blindness of doing exercises of students.In addition,it provided a platform for teachers to understand the dynamic of student and classes error themes. Data analysis could help teachers to understand the strengths and weaknesses of each class and can timely adjust teaching strategies.This paper firstly studied on the background of areas related to the topic and the current situation, secondly, it dug out students’ weak knowledge points by using data mining techniques and it put forward a question recommendation model based on user ratings in view of the weak points,then it proposed an improved K-Means algorithm based on the calculation of classification properties dissimilarity in the process of test recommendation, finally, it designed and implemented the mining system of middle school students error test.The article mined the relationships of students’ weak point of knowledge according to students’ error themes by using Apriori algorithm, and recommend related questions by the relationships for students, and by calculating the similarity between questions it clustered the recommended examination and error test by K-Means algorithm. combining score of the error theme evaluated by the student,it predicted score of recommendation of examination and recommended the tests of higher score to the students. In this paper, the K-Means algorithm was improved,so that the K-Means algorithm could deal with numerical attributes and type attributes.And it proposed a formula for calculating dissimilarity of type attributes.The new formula had better clustering accuracy compared with hemingway distance formulas,so the accuracy of recommendation had been enhanced to a certain degree.This paper set up online error book of students and class based on knowledge points and score weights. It put forward an improved K-Means algorithm based on the calculation of classification properties dissimilarity.This system tailored to students a set of personalized examination questions by mining the wrong weak point of knowledge and using the question recommendation model based on user ratings,thus it strengthened the students’ training of weak points and reduced the blindness of doing exercises of students.