Research Of Reference Automatic Recommendation Based On Data Mining

With the rapid development of computer and network technology, reference resources are increasing very quickly. But the search technology can't meet user's personalized needs to references. On the other hand, teachers take a lot of time to provide references for students in the dissertation step. Under the background, reference recommendation system emerges as the times' requirement. Reference automatic recommendation is used to meet users' interest, and extract their potential preference and provide personalized services to them initiatively.However, cold start, inaccurate match and other factors exist in traditional recommendation method. Data Mining is extracting potential and valuable knowledge (model or rules) from a lot of data. The reference automatic recommendation in the dissertation process is implemented in this dissertation based on DM technology.In the dissertation, the paper present a recommendation method with association rule and classification rule based on the former recommendation methods. The user model is designed, which acquires users' preference by the registration information and the imprint online. The association rule is used to analyze the relationship between reference resources, without sparsity and odd matching problems. Based on the classification rule, users are divided into different groups. Studying users' interest feature helps to recommend reference resources accurately to users with similar preference.In the end, an automatic reference recommendation module based on data mining algorithm is designed and implemented. Experimental results show that the module has a good effect on recommendation in dissertation progress.