论文部分内容阅读
【目的】微博用户兴趣发现对微博社交网络的个性化推荐和提升用户满意度具有重要的意义和价值。【方法】不仅通过挖掘用户自身微博数据识别出用户兴趣,而且进一步挖掘其关注用户的微博数据以及他们之间的社交联系,并通过计算用户微博与其关注用户兴趣的相似度以及用户与其关注用户间的亲密度,进一步发现用户兴趣。最后将从两方面发现的兴趣进行合并,得出用户的兴趣。【结果】基于爬取的新浪微博数据集进行实验,准确率和召回率较传统的方法提升15%以上。【局限】数据预处理中,停用词表不充分,没有实现停用词表的自动学习;需人工标注用户兴趣集计算准确率和召回率。【结论】实验结果表明,该方法明显优于传统方法,能够更加有效和准确地发现用户兴趣。
【Objective】 Weibo user’s interest discovery has important meaning and value to personal recommendation of microblogging social network and user satisfaction improvement. [Method] This method not only identifies the user’s interests by mining the user’s own Weibo data, but also further digs the Weibo data of the concerned users and their social connections. By calculating the similarity between the user’s Weibo and its interested user’s interest and the user’s Concerned about the user’s intimacy, to further discover the user interest. Finally, the interest found in the two aspects will be merged to obtain the user’s interest. 【Result】 Based on the experiment of crawling Sina Weibo dataset, the accuracy and recall were improved by 15% compared with the traditional method. Limitations In the data preprocessing, the stoplist is not sufficient, and the automatic learning of the stoplist is not implemented. The accuracy and the recall rate of the user’s interest set need to be manually annotated. 【Conclusion】 The experimental results show that this method is obviously superior to the traditional method, which can find user’s interest more effectively and accurately.