Study On Adaptive Learning Recommendation Method Based On HIN

In recent years,with the development of the study analysis technology,learners get personalized features and quantitative has formed a more mature theoretical system and practice process,a personalized study plan and study path have been studied.In this case,the adaptive learning system has been developed from theoretical research to practical application.But is not satisfactory,the current recommended results of the adaptive learning system is still not get bigger promotion from the essence,it is also affect the adaptive learning system research and development of the biggest problems.In previous researches,the recommendation functions of adaptive learning system were mostly based on traditional recommendation algorithms based on collaborative filtering recommendation,content-based recommendation and association rule recommendation.Although has obtained certain recommendations,but this kind of algorithm to the learner or the attribute of knowledge unit as the calculation of similarity criteria,ignoring the learners with knowledge unit between the correlation information.However,the analysis and research of heterogeneous network have made a breakthrough in the diversity of data types today.Adaptive learning systems involve multiple classes of objects,and there are different interactions between different objects and the same objects.The object attribute dimension of adaptive learning system is large,and the network composed of frequent connections among objects is a typical heterogeneous information network.So,traditional recommendation algorithm with attributes as the calculation of similarity is not proper,and from the perspective of semantic distance network optimization recommendation algorithm is more in line with the recommendation of the adaptive learning system logic.Aiming at the defect of traditional recommendation algorithm in adaptive learning system,this paper proposes an adaptive learning recommendation algorithm based on heterogeneous information network from the perspective of semantic similarity.In order to improve the effect of adaptive learning system overall recommendation,this paper made a research from the following several aspects:(1)from the Angle of heterogeneous information network network model of adaptive learning system,meta path are studied,and design the research framework based on adaptive learning HIN recommended;(2)based on the learner model specification and the individualized characteristics of learners,the learner model was constructed by using UML modeling technique;(3)the hierarchical structure of domain knowledge was sorted out based on previous studies,and the transformation relationship between knowledge was described by using the bayesian knowledge tracking model;(4)using the similarity measure based on yuan path,design the adaptive learning resources recommendation algorithm based on HIN,and combined with the case of curriculum evaluation and empirical analysis,to determine the effectiveness of the algorithm.