中文摘要：

提出基于链路结构的微博领域专家识别算法，以新浪微博为例，通过多指标综合分析，设计了包括原创发博率、主题相关度、节点扩散度、交互主动性以及用户支持度等五个指标值，将节点的权重分为中心权重和权威权重两部分，分别采用上述指标值为其赋值，再根据用户的中心性和权威性相互增强的思想，迭代计算每个用户的领域中心值和领域权威值，直至算法收敛，最终利用领域权威值结果判断用户是否为领域专家。实验表明，本算法能准确地识别微博中某一领域的专家，并能对其影响力作出有效评判。

英文摘要：

The explosion of the data in network puts forward a new challenge for Web mining. Recognition algorithm of Weibo domain experts is based on the structure of network in Sina weibo, uses the principle of multi-index comprehensive analysis to design five index formulations, including original microblog ration, topic relevance, node proliferation degree, interaction initiative and support degree between users. Then, we divide weight of nodes into two parts, hub weight and authority weight. The above indexes are used respectively to assign to hub weight and authority weight. According to the idea that authority value and hub value will enhance each other, our algorithm iteratively calculation authority value and hub value for each user until convergence. This algorithm has been proved by real dataset from Sina Weibo, and gets an excellent result.