Research On Personalized Modeling Method Of Online Education Learners Based On Multi-source Information Fusion

In recent years,with the integration of Internet + education,various online education platforms have developed rapidly.These platforms have accumulated a large number of users by virtue of their high-quality and massive resources.Online education has become an important educational mode and technical approach for learners to acquire knowledge,develop skills and educate for academic qualifications.How to provide personalized content for learners in the massive educational resources is a hot research issue.The learner personalization m odel records the learner personalization characteristics and can reflect the learner personalization differences.It is an important decision basis for the adaptive learning system to provide the learner personalization services.There have been many researches on learner models,from pedagogy to computer,many scholars have made contributions from various aspects.The relevant researches on learner models in China mainly focus on theoretical research.The factors contained in learner models are comprehensi ve and involve all aspects,but it is difficult to apply them in practical development and their operability is not strong.But the foreign research focuses on and pays attention to a certain characteristic of learners and carries on the practice developme nt and the application research.At present,most learner models are not systematic,their eigenvalues are not comprehensive,their acquisition methods are subjective,their computability is poor,and their researches are mostly at the theoretical level.Therefore,based on the behavioral data of learners on the online education platform,this paper proposes a learner personalization model with multi-source information fusion.From the perspective of learning interest and knowledge background,the representation,extraction and application of learner interest model and knowledge model are studied.Secondly,the multi-level interest model representation method based on knowledge topic,course and domain information is studied.On the real data,based on learners' interests,experiments are carried out for learners' personalized course recommendation,and the effectiveness of interest modeling method is verified.Then,the representation of domain knowledge model is studied and combined with domain knowledge model,the learner's knowledge model can be defined into two parts,knowledge structure and knowledge level.The validity of the proposed knowledge level estimation method is verified by experiments on real data sets.Finally,a complete framework from learner portrait to personalized recommendation is realized.The theory of interest model is applied to practice,and the visual display of the fields,courses and topics that learners are interested in is carried out,and three visualization methods of learner s' knowledge background and knowledge level are realized.