Research On Learning Resources Personalization Recommender Method Based On Web Log Mining Technique

With the rapid development of network technology and online education, more and morepeople acquire knowledge through online learning. The traditional "in order to teach" onlineeducation mode only provides a single teaching resources and does not fully consider thedifferences between individuals. A contradiction generated between Static learning resourcesand the increasing needs of learner’s personalized learning, in order to resolve thiscontradiction, the promotion of personalized teaching system based-on network has becomean inevitable trend. Personalized network teaching system introduced the personalized ideason the basis of the traditional online education. In the system, learners have a lot of freedomand initiative in the learning process; according to the learners’ level of knowledge,educational level, personal learning interests and learning needs, combined with the currentknowledge system and learning ability of the learner, the system recommends learningresources to learners particular and provides learning advice and guidance that meet learner’spersonal circumstances, these can inspire the enthusiasm of the learners, enable them able toindependent and efficient learning and make them achieve the best learning results.Humanistic learning theory, constructivist learning theory and theory of multipleintelligences are the basis theoretical of network-based personalized teaching, they providebasis theoretical for individualized learning. The application of data mining technology inlearning resources recommendation provides an effective way to the implementationindividualized teaching. Using data mining technology, we can extract valuable and regularityinformation from vast amounts of data. Web log data has been the focus objects of datamining. The development of data mining techniques especially web log mining technologyprovides a powerful technical support for the implementation of individualized teachingsystem based on network.This article first describes the background and significance of learning resourcespersonalization recommender method，then analyzes the researches in this field on the current situation at home and abroad in order to understanding the theory and technology ofpersonalization learning. In this work，according to the characteristics of personalized learning,we advantage web log of learning website by fuzzy clustering and association rule mining.Fuzzy clustering is based on the frequency of learning resources accessed at the same time inthe user session. We classify the high similarity of the learning resources to the same cluster，learners will focus on the same classified resource within a period of time. Therefore clustercould narrow the scope of the recommendation resources, and increase the recommendationcoverage. Association rule mining refers to hide the relationship between the data by using ofweighted association rule mining, we closely the recommendation of related resources inorder to improve the recommendation accuracy，We use the weighted association rules byimproving data mining apriori algorithm.After selecting log clustering method and data association rules, the paper researches theconstruction of learning resources personalized recommendation model and evaluate theperformance of the system using the evaluation parameters of personalized recommendation.First, we research on learning resources personalized recommendation model based on weblog mining, and then describe the implementation of learning resources personalizedrecommendation mechanism, analyze the evaluation parameters of the system finally. In orderto verify the validity of the algorithm and model indicators, we do some experiments usingthe learner access log of Chinese learning network as the data set, and the result shows thatthe algorithm proposed in the article improves the precision and coverage of therecommendation of learning resources.The learning resources personalized recommendation model implemented in this articlecan recommend learning resources to the learner accurately, but there are still someinadequacies. In future work, we have to constantly improve the model to make the userexperience better.