| Collaborative filtering recommendation algorithm is a most widely used algorithm in personalized recommendation field today. A microblogging reference and user features&category based algorithm is proposed. The algorithm joins the reference category and user features similarity to user’s similarity calculating. First, calculate the users’ personal information and language habits similarity on multi-demensional, then calculate the users’similarity of the feature of category s’rates based on microblogging reference contents. Then conbine the similarities weightly, calculating the integrated similarity. Get a close neighbor set of target user according to the integrated similarity. Then sort the close neighbor set by similarity to get a result of recommended microblogging users set of target user. Predicting the target user preferences microblogging item score according to the close neighbor set’s weighted score of the microblogging item. Finally generate the recommended microblogging result by predicted result.In order to select the. computing users’set of microblogging personalized recommendation, a data extraction method based on users’ relationship and category feature is proposed. Thus simplifying the number of computing set greatly, and the performance of algorithm has been optimized greatly.This paper selects the data set of SINA micro blog as the experimental data, in order to verify the effectiveness of the algorithm in this paper. The method is, through the different comparative selection of threshold and weights of the collaborative filtering recommend algorithm, which will influence the recommended results, and accomplish the analysis of the experimental results and assess the performance of the designed system. The experimental results suggest that the proposed algorithm has a good performance and it can achieve the personalized recommendation based on the target user’s interests from sina micro blog, such as some hot micro-blogging and famous users. |
PDF Full Text Request