Study On Intelligent Recommendation For Learner Leading Course Based On Rule Base

The Education Informatization 2.0 Action Plan was issued by the Ministry of Education in April 2018,which means that network education has risen to the national strategic position.However,because of the mismatch between learners' knowledge and curriculum,the problem of "high dropout rate and low completion rate" has seriously affected the development of online education.There are three basic reasons:(1)the existence of advanced relations among courses;(2)the blindness on learners' curriculum choices;(3)few guidance services for leading courses on the platform.From the literature review on the previous research,it is obvious that there exists the mismatch between learners' knowledge and curriculum,though the problem of "overload" on curriculum information has been solved.Therefore,it is of great importance for learners to make the progress on curriculum and avoid choosing courses blindly.Therefore,this paper aims at providing intelligent recommendation service for leading courses and taking learners and their selected advanced courses as the object,based on learning advanced theory and the data carrier on the Icourse163 to solve the mismatch between knowledge reserve and curriculum.It involves in three aspects including rule base construction,learners' health perception and intelligent recommendation for leading courses.The innovation of this paper were as follows:(1)Based on the advanced relations of curriculum,a rule database is constructed.With the help of the advanced learning theory,web crawler technology and text analysis method,the research is used to select and standardize the advanced relations among the platform courses as well as construct the rule base.The construction of the rule base makes the relations among 1744 leading courses clear.(2)Based on the rule base,a perception model for health of course selection is constructed.The learner behavior characteristics are collected and analyzed in terms of Icourse163,the relevant data obtained by web crawler technology and the statistical analysis.According to learner's selected leading courses,the prediction method of course mastery is proposed in the field of Logistic function;while as for unselected leading courses,the prediction method of course choosing is given in collaborative filtering algorithm.Besides,in the use of piecewise function,the combination between these two prediction methods is realized.It is meaningful to construct this perception model so as to predict the condition of learners' knowledge.(3)An intelligent recommendation method for learner-led course is proposed.Based on the learners' health perception model,the research realizes an intelligent recommendation method for learner-led courses.This method uses rule base to perceive learners' healthy condition and then recommends the required leading courses for learners in terms of the perceived results and recommendation rules.The results show as the number of recommendations is within1 ?n ?5,the average accuracy is 9.16%,which demonstrates that the proposed method has good perception and recommendation of the leading courses.There are two significance for this paper.On the one hand,the research enriches the advanced theory for learning by the study on curriculum advancement.On the other hand,the research solves the problem of the mismatch between learners' knowledge and courses by the study on the blindness of learners' course selection.