Research On Course Recommendation And Learning Forecasting Model Based On Deep Learning

With the rapid development of the Internet,Internet learning community has been deep into the lives of the majority of Internet users.It is constantly affecting the Internet user's learning at an amazing rate.At the same time,Internet learning community produces knowledge overload,learning trek and other new learning problems.Learning resources recommendation can be used to solve such problems,and it is also an important means for learners to learn quickly and efficiently.Therefore,researching and developing the system of learning resources recommendation which determines the type of learners by analyzing learning records,learning preferences and other information of learners and recommends learning resources that meet the learners' interest,has become a hot topic in the field of Internet learning and education.The main works of this paper are follows:(1)This paper summarized the development of Internet learning resources and the future development trend.At the same time,it elaborated some main technologies and methods in the personalized recommendation system,and made a deep explanation of the principle of personalized recommendation technology.(2)This paper proposed the course recommendation model which adds the user information and course information into the process of potential user feature training and potential course feature training based on the NeuMF model.This method is used to train the model by adding information such as user gender,occupation,course orientation,course classification and so on and combining with user ID and course ID.In order to verify the effectiveness of the model trained in this way,this paper used the actual data set from iMOOC to compare the experiment.The result shows that the training model obtained by this method has better performance than the original model.(3)This paper proposed a learning prediction model based on convolution neural network,which integrates the user information and the course information into the implicit factor decomposition model.The method extracts the user feature information from the user ID and the user's personal information,extracts the course feature information from the course ID and the course classification through the convolution neural network,and constructs the relation vector describing the user and the course relationship.Finally,it obtains the final prediction result through computing the relation vector.In order to verify the validity of the model,this paper used the real data of iMOOC to carry out the comparison experiment with other models.The experiment result shows that the method has better predictive results.