【Abstracts】 Using association rules to mine the college library circulation information, you can find out the borrowing habits of readers to provide readers with the books they are interested in. However, in the actual usage, the statistics showed that the recommended mode based on the traditional association rules has some problems that it's difficult to find association rules and it doesn't reflect the change of the readers' reading habits. This paper attempts to recommend a higher hit ratio personalized recommendation model to users by analyzing the dugout results and giving a special weighting disposal. Actual tests show that the association rules given a special weighting disposal make the recommended model to reflect the users' changes in time and are able to meet the need that provides the users with a more professional personalized recommendation.
【Keywords】 Association rules; Individuation; Recommendation model; Data mining; University library