Research Of Decision Tree Classification And Its Application To Teaching Analyze

The Data Mining is also called the knowledge detection from the database, which is the process of extracting valid, implicit, unknown, latent information from the large scale of database. It is one of the research topics in many scientific areas too, especially in the database realm. The research of Data Mining has reached significant achievement and has been applied successfully in many areas. However, successful application of Data Mining in the field of education has not been reported.With more and more scientific ways used in the students management and teaching practice, the traditional means have already can't fulfill the need of the social developments. But if we use the successive grade database of students as data source, and make use of the data mining techniques to discover how teacher's, student's and textbook's informations influence the student's grade, then we can establish the reasonable teaching projects according to the above data mining results.Under general situation, the algorithm of Data Mining can be divided into the classification and prediction, the gather kinds, the analysis and the connection rules mining...etc. Among those, the classification on decision trees is one of the most important and common techniques, which is also a hotspot study currently.According to the above conditions, the article has done the following research:1. Research the various decision tree algorithms and analyze the main technique characteristics of each algorithm.2. Based on the relationships of students' grade and teaching request, the choice of textbooks, the circumstance of the teachers and the students, the author make use of the traditional realization method of ID3 to carry out the grade analytical system.3. While analyzing the traditional realization method's defect carefully, this article put forward an improved realization method CASP (Classifying Algorithm on Server Procedure), which is also applied to the grade analytical system. Comparing with the traditional realization method, through the theory analysis and the actual data tests, the improved realization method CASP is proved to reduce the memory and calculation complications.In the grade analytical system design, the author adopt the model of MVC( Model- View- Controller) and use VB.NET and SQL Server 2000 to realize. The system is divided into two parts: data training and data predicting. The training part is used to clean the teaching history datas (training recordset), disperse the consecutive datas and then form the decision trees and the rules through the decision trees algorithm .The predicting part is to test the current datas by the rules, give a visual interface for the test and predict results. Through the practice, this system can provide the important basis for teaching analysis and curriculum arrangement.