Research Of Association Rule In Analysis Of Students' Score

With the popularization of higher education, continuous enrollment of each college makes the number of students increase dramatically. It brings us not only universal education, but also the rapid growth of information which results in a lot of serious education problems for the management of school administrators and teachers. Traditional education management methods cann't satisfy the education and management work under the new situation. In the face of these huge data, the traditional method has appeared to be inadequate. How to manage students' scores scientificly and effectively is an urgent problem. Data mining can be an effective solution to this problem.Most colleges and universities have their owner student score database, but only for some simple query and statistical operations. Some vluable information behind these score data is not used. To solve this problem, this paper proposes to use data mining methods on student score data to get some hide information between different courses in order to facilitate decision-making analysis of school administrators, teachers and students. The main work contents and achievements are as follows:1,The traditional association mining algorithm FP-growth has low efficience. To solve the problem, this paper proposes the FPT algorithm to improve its data structure with addition of a tail domain, organize all items to generate FPT-tree. When adding a node, the traditional method requires the first node from the list to find the beginning, and then follow the next pointer of each node, until you find the last node, then add the new node into the list tail. FPT FPT-tree algorithm to find the tail domain directly linked list can be found in the last node, you can directly insert the new node to improve efficiency. Experiments show that the improved FPT algorithm has improved greatly in performance, especially in processing large amounts of data, this advantage is more apparent.2,Because the student score is a discrete numeric data, in order to user association mining algorithms, the range of student score need to be divided. However, the size of the interval length has a direct impact on the generation of association rules. In order to minimize this effect, the introduction of Bulr Set associated with this mining, blur association mining algorithm proposed called B-Apriori. Set of properties will need to tap into a value of Bulr Sets, Bulr Sets its data item calculated the degree of membership, by membership of the item set to calculate the support and confidence. Experiments show that the B-Apriori algorithm has high accuracy and practicability.3,Design and implement the student score management and analysis system based on FPT algorithms and B-Apriori algorithm to find the association between different courses which can privode the decision-making for teachers and can be also useful for students.