Design And Implementation Of Online Learning Resource Recommendation System Based On IEUS-Bagging

With the rapid development of Internet and education,online learning has gradually become a hot field,followed by the redundancy of online learning resources,learners can not quickly obtain the desired learning resources,which directly affects learners' interest and patience.Therefore,in the era of information overload,it is of great significance for learners to mine and accurately push learning resources that meet their immediate needs.In order to classify learners' learning level more accurately and provide learning resources that meet their own needs,this thesis uses Bagging integration algorithm to analyze online learning behavior and classify learners' learning level according to the unbalanced characteristics of online learning behavior data.At the same time,considering the existing problems of Bagging integration algorithm and make improvements.In this thesis,the improved Bagging ensemble algorithm is applied to online learning resource recommendation system,the main work of this thesis is as follows:(1)A Bagging ensemble algorithm based on information entropy is proposed.Aiming at the problem that bagging ensemble algorithm is easy to cause the loss of important sample information in the unbalanced data set when sampling,this thesis proposes an undersampling method based on information entropy,and uses the error sharing rate combined with log function weighted voting method to solve the problem that the weight of individual learners is the same,which easily leads to the poor performance of the algorithm.Finally,combining the two methods,this thesis proposes the IEUS-bagging ensemble algorithm based on information entropy under sampling optimization(IEUS-bagging algorithm for short),and describes the process of IEUs bagging algorithm.Then,the effectiveness of IEUS-Bagging algorithm is verified by experiments.(2)A hybrid recommendation model of online learning resources based on IEUS-Bagging algorithm is proposed.Firstly,according to the needs of online learning behavior analysis,the learner model and online learning behavior analysis model are established.Finally,three different recommendation models are constructed from the three dimensions of learners' cognition,interest and learning behavior.Before recommendation,IEUS-Bagging algorithm is applied to the classification of learners' learning grades,and an online learning model based on IEUS-Bagging algorithm is proposed.The experimental results show that the classification method based on IEUS-Bagging algorithm is effective in imbalanced learning behavior dataset.(3)Completed the design and implementation of online learning resource recommendation system.Aiming at the needs of online learning resource recommendation system,this thesis applies the hybrid recommendation model of online learning resources based on IEUS-Bagging algorithm to the recommendation system.By using the existing development technology to design the system to achieve the online learning resource recommendation system and put it into use.Finally,the application data of the recommendation system is analyzed to prove that the system is of practical significance.