Based On Decision Tree Network Of College Students’ Academic Research Influence Factors

Since the Ministry of education began to Pilot run remote education energetically in1999,network academy play in his positive role also has exposed some disadvantages: due to the complexity ofsource of learners of network education, the environment is not fixed, enough makes effort and learningtime cannot guarantee, that dropping out of school, hysteresis phenomenon become a serious problemfacing network educators.At present, the influence factor of network education school research is still in thelevel of education statistics,, many research are still based on a low level such as experience and feelingsetc. Empirical research method is not enough scientific and standardized, such as questionnaire design,sampling representativeness, lack of strict data statistical reasoning, etc, questionnaire design, samplingrepresentativeness, lack of strict data statistical reasoning, etc.In this paper, from several aspects to analyze Network Academy without finish school and geteasy decision model: first of all, clear the research background of this study, the research present situation,the purpose of the study, the structure of the paper, etc.; Secondly, describes the development of our countrynetwork higher education present situation, and through the literature research and theoretical analysis,understand the "adult online learning","data mining" and "decision tree" and other relevant concepts andtechniques;Then improved Kember model, put forward to adapt to the characteristics of "network educationthrough school influence factors model", and to design questionnaire, get information about studentlearning process; Finally combined with the information in the distance learning system composition to thedata objects, we introduce the decision tree technology, using the Weka data mining software to carry onthe analysis, find out the influence degree of each factor for network learners to complete their studies, andby using the results of the analysis may have dropped out of a dangerous new entrants to implementcorresponding countermeasure, have a purpose to give guidance to improve the early warning conversionrate.The main task of this topic is according to the location of a remote education college students’network learning behavior information, using data mining technology to analyze the relationship betweenthem and completion of, and concluded that an optimal decision tree rules based on the academic influence factors. Through data analysis, the most worthy of our alert to the influence of network education academicfactors, Data analysis model, created by institute of education policymakers and new entrants to learners ofreading intervention management provides a powerful means and it is of great significance to improve thenetwork higher education quality.