[目的/意义] 微博转发是实现微博信息传播的重要方式，对用户转发行为进行研究可以更好地理解微博信息传播机制，对热点话题检测、舆情监控、微博营销等具有重要意义。针对以往研究中用户兴趣表示不够全面准确以及未考虑情感差异对用户转发行为的影响，提出一个融入情感差异和用户兴趣的微博转发预测模型。[方法/过程] 该模型首先从维基百科中提取概念语义关系构建维基知识库，将其作为语义知识源对微博文本进行语义扩展，解决语义稀疏问题；对语义扩展后的用户历史微博进行聚类，提取用户兴趣主题和主题对用户的影响力；然后计算微博中各类情感的情感强度，提取情感差异特征；最后结合用户行为特征、用户交互特征、微博特征、用户兴趣特征和情感差异特征，运用SVM实现微博转发预测。[结果/结论] 在新浪微博真实数据集上进行实验，验证了所提模型的有效性。
[Purpose/significance] Microblog retweeting is the key way for information diffusion. The study of user retweeting behavior can better understand the information diffusion mechanism in microblog, which is of great significance for hot topic detection, public opinion monitoring and microblog marketing and so on. In this paper, we analyze the factors that affect users' retweeting behavior. The users' interest isn't comprehensive and accurate, and the influence of emotional divergence on users' is not considered in the previous research. Thus, we propose a microblog retweeting prediction model integrating emotional divergence and user interests.[Method/process] Firstly, we built a Wikipedia knowledge base by extracting semantic relations from Wikipedia, and extended the semantic feature of microblog text vector by using the Wikipedia knowledge base, which could solve the semantic sparse problem. To extract users' interest themes and their influence on users, we clustered the users' microblogs which was extended by Wikipedia knowledge base. Secondly, we calculated the emotional intensity of all kinds of emotions in microblog, extracting the emotional divergence features. Finally, combining users' behavior features, users' interaction features, microblog features, users' interests features and emotional divergence features, we used SVM to achieve the prediction of microblog retweeting.[Result/conclusion]The experimental results show that the proposed method can effectively improve the performance of microblog retweeting prediction.
[1] 张亚明, 唐朝生, 李伟钢. 微博机制和转发预测研究[J]. 情报学报, 2013, 32(8):868-876.
[2] PFITZNER R, GARAS A, SCHWEITZER F. Emotional divergence influences information spreading in Twitter[C]//International AAAI conference on weblogs and social media. Menlo Park:AAAI Press,2012:2-5.
[3] BOYD D, GOLDER S, LOTAN G. Tweet, tweet, retweet:conversational aspects of retweeting on Twitter[C].//Hawaii international conference on system sciences. Kauai:IEEE,2010:1-10.
[4] SUH B, HONG L, PIROLLI P, et al. Want to be retweeted? Large scale analytics on factors impacting retweet in Twitter network[C]//IEEE second international conference on social computing. Minneapolis:IEEE, 2010:177-184.
[5] ZHAO X, ZHU F, QIAN W, et al. Impact of multimedia in Sina Weibo:popularity and life span[C]//Joint conference of 6th Chinese semantic web symposium and the first Chinese web science conference(CSWS & CWSC). Resewch collection school of information system. New York:Springer,2013:55-65.
[6] BIAN J, YANG Y, CHUA TS. Predicting trending messages and diffusion participants in microblogging network[C]//Proceedings of the 37th international ACM SIGIR conference on research & development in information retrieval.New York:ACM Press,2014:537-546.
[7] 张玢玢, 李兵, 李岳欣. 基于特征选择的企业微博转发机制研究[J]. 情报杂志, 2014(12):127-132.
[8] 张旸, 路荣, 杨青. 微博客中转发行为的预测研究[J]. 中文信息学报, 2012, 26(4):109-114.
[9] 曹玖新, 吴江林, 石伟,等. 新浪微博网信息传播分析与预测[J]. 计算机学报, 2014(4):779-790.
[10] 刘玮, 贺敏, 王丽宏,等. 基于用户行为特征的微博转发预测研究[J]. 计算机学报, 2016, 39(10):1992-2006.
[11] 陈江, 刘玮, 巢文涵,等. 融合热点话题的微博转发预测研究[J]. 中文信息学报, 2015, 29(6):150-158.
[12] 谢婧, 刘功申, 苏波,等. 社交网络中的用户转发行为预测[J]. 上海交通大学学报, 2013, 47(4):584-588.
[13] 吴凯, 季新生, 刘彩霞. 基于行为预测的微博网络信息传播建模[J]. 计算机应用研究, 2013, 30(6):1809-1812.
[14] 李志清. 基于LDA主题特征的微博转发预测[J]. 情报杂志, 2015(9):158-162.
[15] STIEGLITZ S, DANG-XUAN L. Emotions and information diffusion in social media-sentiment of microblogs and sharing behavior[J]. Journal of management information systems, 2014, 29(4):217-248.
[16] KANAVOS A, PERIKOS I, VIKATOS P, et al. Modeling retweet diffusion using emotional content[M].Artificial intelligence applications and innovations. Berlin:Springer,2014:101-110.
[17] NAVEED N, GOTTRON T, KUNEGIS J, et al. Bad news travel fast:a content-based analysis of interestingness on Twitter[C].//Proceeding of 3rd ACM WebSci conference. Germany:Web Science,2011:1-7.
[18] Index of/zhwiki/latest/[EB/OL].[2016-10-20].https://dumps.wikimedia.org/zhwiki/latest/.
[19] DKPro JWPL[EB/OL].[2016-11-12].https://dkpro.github.io/dkpro-jwpl/.
[20] GABRILOVICH E, MARKOVITCH S. Wikipedia-based semantic interpretation for natural language processing[J]. Journal of artificial intelligence research, 2014, 34(4):443-498.
[21] HUANG J, CHENG X Q, SHEN H W, et al. Exploring social influence via posterior effect of word-of-mouth recommendations[C]//International conference on web search and web data mining. New York:ACM Press, 2012:573-582.
[22] 徐琳宏, 林鸿飞, 潘宇,等.情感词汇本体的构造[J]. 情报学报, 2008, 27(2):180-185.
[23] CHANG C C, LIN J. LIBSVM——a library for support vector machine[EB/OL].[2016-10-24].s.http://www.csie.ntu.edu.tw/~cjlin/libsvm/.
[24] 八爪鱼采集器[EB/OL].[2016-09-20].http://www.bazhuayu.com/.
[25] 停用词集合(哈工大停用词表、四川大学机器智能实验室停用词库、百度停用词表)[EB/OL].[2016-10-12].http://www.datatang.com/data/19300.
