Research Of Online Education Students’ Academic Achievements Subdivide Prediction

Nowadays, a variety of learning and management systems for network education are becoming perfect increasingly. Formative assessment systems of network learning are prompting more and more online learning behaviors really happen. Learning through the Internet, students amassed a lot of useful data and information in learning and management systems. In this research, having analyzed the English learning and some other relative information of undergraduate students in network education, we achieved the subdivision prediction of their results in the English examination for adult bachelor degree, though a general runtime environment for data mining-SPSS Clementine12.0. First of all, we used the K-means algorithm to determine the interval distribution of well chosen existing students’ scores in the examination. Then, setting this distribution as the predictive aim, we employed the C5.0 algorithm to structure the subdivision predicting rule for the English examination for adult bachelor degree, and, we proposed implementation method of the predicting rule and realized it. Additional, we discussed the importance of all dependent variables which constructing the rule, and put forward the corresponding strategies of improving the learning levels and English examination results in the adult bachelor degree for undergraduate students in network education.The raw data this paper used is the 15000 data from 2009 to 2013 of the Network Education Institute of Jiangnan University,after data cleaning we use 8000 data to construct degree English achievement subdivision prediction rule and develop the corresponding system.Use the data mining tool Clementine12.0 to construct the degree English achievement subdivision prediction rule and develop this system to realize the network education of adult undergraduate students’ English scores is the focus of this paper.First, only retain the attribute about degree English performance,cluster analysis on it to determine the specific performance distribution interval which is the classification target. Then use the classification algorithm to construct the prediction rule by 8000 data to program to realize the degree English performance subdivision prediction system which can be carried out to forecast the degree English achievement of adult undergraduate students on Network education,and to provide early warning and learning suggestions according to the forecast results. Finally,to verify the effectiveness of the system by comparativing and analysising the system performance and the actual results and questionnaire and interview,and questionnaire and interview this study.