Research And Application Of Online Education System Supporting Recommendation

With the rapid development of information technology,people have already become information overload from the information deficiency.Based on users' behavior,the recommendation system can quickly recommend valuable information for users from tons of data,so that it can shorten search time and improve users' experience.SAP is a popular software for providing business solutions and applications,and it is developing in China.But there is no MOOC website for SAP so far.According to mentioned above,a novel recommendation algorithm is proposed and online learning website is implemented in this paper.As for algorithm design,we use a open source segmentation tool to extract words from text and calculate the importance of words using TFIDF.Then cosine formula is used to calculate the similarity of courses.Based on Item-CF,the similarity of courses is calculated combing with the courses score.Two algorithms is combined by weighted method.Finally,the list of Top-N courses with the highest similarity is recommended to the user.As for system design,SAP e-learning system is designed and implemented using MVC architecture.The system includes functions such as login,registration,searching,recommendation etc.The recommendation module combines scores and text to recommend courses using the proposed algorithm.The interactions is implemented between website front-end and database.The MySQL database is selected to design the key form.At the same time,the solutions of cold start problem is gave for user and courses.In summary,the SAP e-learning recommendation system is implemented by MVC.Database creation,system interactions and recommendation using scores and text are realized.The off-line experiment is adopted to analyze performance of the proposed algorithm,compared with two existed algorithms Item-CF and Tags.The recommended course is calculated according to the algorithm proposed.The results show that the proposed algorithm is better than Item-CF and Tags in terms of accuracy and recall performance.