基于监督随机游走的有影响力用户发现算法

基于用户主题偏好的监督随机游走是当前热点研究之一。受微博文本稀疏性的影响，基于微博主题衡量用户影响力的研究较少。为研究某些领域的用户影响力强弱，文章提出一种基于监督随机游走的有影响力用户发现算法(TP-SRW)。通过研究微博用户自身的属性和微博文本的主题特征，构建基于用户交互信息的个体链接关系，并结合代表用户同质性的微博主题兴趣的相似性，构建有向加权网络模型。实验结果分析表明，在不同主题下挖掘高影响力用户，提出的TP-SRW算法和模型具有良好的性能。

An Influential User Discovery Method Based on Supervised Random Walk

Now, it is a hot topic to research supervised random walk based on user topic preference. Due to the text sparseness of the microblog, less and less research results have been conducted on the use of microblog topics to measure user influence. To evaluate the influence of users in certain fields, the paper proposes a topic preferences supervised random walks (TP-SRW) algorithm. By use the microblog users' own attributes and the microblog text topic features, the TP-SRW algorithm constructs user-link relationships based on user interaction information, combines the similarity of the interest of the microblog topics that represents the homogeneity of the users and constructs directed weighted networks model. Finally, the experimental results show that the TP-SRW algorithm is better than others in terms of calculating users influence.