The Application Of Data Association Mining And Its Improved Technology On The System Of The Evaluation Of Students

At present, there are numerous students study in the colleges and universities. And, the rewards and allowance that government and universities give to the students are more and more. To handle such kind of work, like the comparison and appraisal of students by handwork is easy to make mistakes and inefficiency, the comparison data also cannot be kept totally. Therefore, a management information system (MIS) is urgently needed to process all kinds of comparison and to keep the total comparison and appraisal data. At the same time, it is also urgently to find out useful rules from these data to guide the follow-up comparison work, and compare and appraise automatically. Aiming at solving this problem, this paper adopts the Association Rules Technology of Data Mining to dig out all kinds of comparison rules from the data the system kept. Then, for the poor students’stipend comparison, setting up a classifier according to the finding rules by using the associative classification technique to classify the prospective data automatically. So, the comparison and appraisal work can be done automatically, the efficiency and the fairness also can be improved.In order to solve these disadvantages, like scanning the terms in database of the Association Rules Apriori Algorithm for many times, this paper bring forward the improved Apriori Algorithm to improve the mining efficiency of the frequent terms while finding out the rules from the data the system kept. In the improved Apriori Algorithm, firstly, it will form the upper triangular binomial spaces support matrix by scanning the database one time. It can obtain frequent one-type and frequent binomial, and also can generate candidate trinomial quickly through this matrix. Then, on the basis of the finding candidate trinomial and according to the features of the Apriori Algorithm, it will form a linked list by scanning the data that generating function to the counting support of follow-up candidate terms from the database. The follow-up data mining will be finished in the linked list, and safeguard the linked list dynamically to make every node in the linked list always formed from the terms that generating function to the counting candidate terms support, until it cannot generate longer frequent terms any more.The associative classification of CBA Algorithm has some advantages, such as classifying fast, high accuracy, etc. However, it will generate over many class association rules. Therefore, this paper brings forward the optimization of the CBA Algorithm. For the optimization, firstly, it will use the frequent closed terms to keep the frequent terms that mined. This method can avoid keeping huge frequent terms and decrease the number of generating class association rules. Then, it will use one method that mainly adopt the whole confidence and supplement lift to make sure the generating class association rules are positive correlation. Finally, using some rules from the classifier to predict the classification label of the prospective classification data. These optimization methods can decrease the number of generating class association rules, and improve the accuracy degree of classification at the same time.Last but not the least, this paper adopts SQL Server2008as the background database, and uses JSP technology to implement the students evaluation system. In this system, it uses the improved Apriori Algorithm to mine the comparison and appraisal data in the database, and finds out all kinds of relevant rules. While comparing and appraising the poor students’stipend, it can set up the poor students’classifier according to the data from the system and classify the prospective data. This makes the comparison and appraisal of poor students’ stipend automatically come true.