Research On The Course Recommendation Based On Word2Vec And TF-IDF

With the rapid development of Internet technology,online education provides favorable conditions for learners' personalized learning and lifelong learning.However,with the construction and development of the online education platform,the types of courses are constantly enriched and the number is greatly increased.The problem of information overload is becoming more and more prominent.In addition,the online education platform at this stage mostly provides functions such as classification and search,and there is a problem of insufficient adaptability and personalization in terms of courses.How to quickly locate the courses of interest according to the needs of learners and historical behavior data is an appeal for personalized learning and an important issue to be considered in the construction of online education platforms.This study takes the cadre online education and training platform as an example.In response to the problem of insufficient semantic analysis in the modeling of traditional content-based recommendation algorithms,the Word2 Vec model in the field of natural language processing is introduced,and a course recommendation algorithm based on Word2 Vec and TF-IDF is proposed In order to improve the accuracy of course recommendation,the effectiveness of the algorithm is verified through the data of the cadre network education and training platform,and it is applied to the course recommendation of the platform.The specific work of the study can be summarized as follows:Firstly,based on Word2 Vec and TF-IDF course recommendation algorithm research.The study combs the current status of domestic and international educational resource recommendation,analyzes the common recommendation algorithms and principles in the recommendation system,and designs a curriculum recommendation algorithm based on Word2 Vec and TF-IDF,which combines the advantages of semantic analysis of word vectors and the importance of word frequency statistics Sex.The recommended modules for the course include Word2Vec-based word vector representation,word vector clustering,feature vector representation,and similarity calculation.Secondly,experiment and test the course recommendation algorithm.First,the research standardizes the original data sets obtained from the platform.There are 1539 online courses and 240124 students.Secondly,the course data is used as a corpus for experiments to realize text preprocessing,training word vectors,clustering,similarity calculation and other processes,so as to obtain a course recommendation list.Finally,in order to test the recommended effect of the algorithm,research and design an offline experiment for testing.There are two main testing schemes.First,design 24 sets of comparative experiments to test the recommendation effect of different clustering methods(adaptive clustering,K-mean clustering,Birch clustering)to determine the optimal solution.Second,the recommendation accuracy of different algorithms(mean Word2 Vec,TF-IDF,Word2 Vec,and TF-IDF)was tested to verify the effectiveness of the algorithm.Finally,design and implement a course recommendation system.The research starts from the requirements,designs the system's functional structure,database E-R diagram and data table structure and completes the system development.As the study takes the cadre network education and training platform as an example to conduct algorithm experiments and tests,the curriculum recommendation system built on this basis not only presents the results of the algorithm,but also solves the problem of difficulty in selecting courses for learners.supplement.