Knowledge Map Recommendation Algorithm Based On Interactive Behavior Analysis

Due to the complexity of the structure of knowledge points and the diversity of behavior in the process of online learning,there are differences in the learning knowledge map and learning effect of learners with the same cognitive level in the process of online learning.This paper conducts collaborative analysis based on the multi-dimensional interactive behavior data generated by learners in the online learning platform,and mainly conducts the following research:1.Aiming at the problem that the existing knowledge map recommendation algorithms do not effectively mine the rules of learners’ interaction behavior,and the limitation of similarity calculation in collaborative filtering recommendation,this paper proposes a personalized knowledge map recommendation algorithm based on collaborative filtering.Firstly,the interaction degree of knowledge points is described according to the online multidimensional interaction behavior,and the mastery degree of knowledge points is described according to the online test results,and the learning effect is described combined with the interaction degree and mastery degree of knowledge points.Secondly,the correction factor based on the difference of interaction degree difference of knowledge points is studied.According to the average interaction degree of knowledge points,the weight coefficient of the difference of interaction degree of knowledge points is described,and the similarity calculation model is improved,which effectively avoids the limitations of similarity calculation and improves the prediction effect.Finally,the personalized knowledge map is recommended to the target learners according to the learners’ similarity and learning effect.2.Aiming at the problem that the personalized knowledge map recommendation algorithm based on collaborative filtering does not consider the difference between the target recommendation learner with its similar neighbor and the diversity of recommendation,this paper proposes a multi-objective recommendation knowledge map correction algorithm based on the difference between similar learners.Based on the difference of the initial cognitive level between the target recommendation learners and their similar neighbors,this algorithm proposes four correction rules from the overall structure,local structure and node attributes of the knowledge map,and makes multi-target correction for the recommendation knowledge map based on collaborative filtering.The proposed algorithm is verified According to the mastery degree of knowledge points of target recommended learners.The effectiveness of the proposed algorithm is verified by the accuracy rate,recall rate,F1 value and MAE value.According to the difference between the recommended knowledge map and the actual learning knowledge map in learning duration,learning path and learning effect,the significance of the algorithm in practical learning is analyzed.The experimental results show that the proposed algorithm has achieved better recommendation effect,which is helpful to improve the learners’ online learning effect and learning efficiency.