Research On The Method Of Course Recommendation For Integrating Deep Learning

With the advent of the mobile Internet era,information in the Internet has grown exponentially,and at the same time,the way people access information is more convenient.But the problem that comes with it is that in a lot of information,we usually feel confused when we get it,and we find it longer and longer.The advent of the recommendation system helps users find resources that are appropriate for the user,so that users can pay more attention to the resources of the province rather than spend time looking for resources.With the outbreak of online courses,various courses have moved from the three-foot platform to the screen,and learners have more choices to learn what they want to learn,but since the current research on course recommendation is still in the early stage,this is Due to the complexity of education.Recommending the right course for the learner not only considers the content of the course itself,but also focuses on the learner,such as the emotional state of the learner's learning.In order to solve the problem of accuracy and efficiency of course recommendation,this paper designs a collaborative filtering recommendation method based on doc2 vec and a course recommendation method based on sentiment analysis by integrating deep learning methods from the perspective of curriculum and learners.The research content of this paper is mainly divided into three aspects:(1)Data collection and pre-processing.In this phase,this paper designs a distributed crawler to obtain data in the Internet,and performs operations such as data cleaning and Chinese word segmentation.Finally,the sentiment analysis training and test set are obtained by designing the artificially annotated sentiment analysis data.(2)Design a collaborative filtering recommendation method based on doc2 vec.In this stage,the paper uses the deep learning model doc2 vec to perform document vector training on the acquired text,and combines the collaborative filtering method to recommend the appropriate course for the user.(3)Design a course recommendation method based on sentiment analysis.Based on the convolutional neural network,this paper constructs an sentiment analysis model and obtains a high accuracy rate in the course review text sentiment analysis data set through data set training.Finally,the candidate recommendation set is obtained by combining the neighboring value between the sentiment analysis result and the user.Based on the data obtained,this paper compares the general recommendation model and obtains better recommendation results.The effectiveness of the proposed method is proved by experiments.