微博网络中用户关注行为预测

发布时间：2018-02-11 03:14

本文关键词： 微博用户关注预测模糊近似支持向量机　出处：《系统工程》2015年07期 　论文类型：期刊论文

【摘要】：在微博网络中挖掘博主的潜在关注用户,建立相应的关注预测指标和模型,对于增加博主活跃粉丝数、增强信息传播具有重要意义。利用微博粉丝人际关系网络,在众多与博主交互的用户中,预测潜在的活跃粉丝。在融合微博主影响力、用户活跃度等因素的基础上,提出了以兴趣相似度为核心的用户关注预测指标。利用模糊近似支持向量机作为预测工具,对模糊隶属度算法进行了优化,推导了基于矩阵的模糊近似支持向量机预测模型。利用KDD Cup 2012腾讯微博公开数据进行了实证研究。实验结果表明,本文提出的指标能很好地用于微博用户关注行为的预测,通过与其他支持向量机模型及模糊化算法相比,本文提出的优化方法具有较高的预测精度,且预测结果稳定,在微博用户关注行为预测中具有较强的实用性。
[Abstract]:It is of great significance to excavate potential users of bloggers in Weibo's network and establish corresponding concerned prediction indexes and models to increase the number of active followers and enhance the dissemination of information. Among the many users interacting with bloggers, predict potential active fans. Based on the integration of Weibo's main influence, user activity and other factors, In this paper, the user concerned prediction index with interest similarity as the core is proposed, and the fuzzy membership algorithm is optimized by using fuzzy approximate support vector machine (FASVM) as a prediction tool. The prediction model of fuzzy approximate support vector machine based on matrix is derived. An empirical study is carried out by using KDD Cup 2012 Tencent Weibo open data. The experimental results show that the proposed index can be used to predict the user's concern behavior. Compared with other support vector machine models and fuzzy algorithms, the proposed optimization method has high prediction accuracy and stable prediction results, and has strong practicability in Weibo user attention behavior prediction.
【作者单位】：电子科技大学经济与管理学院;四川农业大学建筑与城乡规划学院;
【基金】：国家自然科学基金资助项目(71101018)
【分类号】：F49

【参考文献】