Microblog user influence ranking method integrating user time information

The invention relates to a microblog user influence ranking method integrating the user time information. The method comprises steps that (1) characterizing microblog users is carried out; (2) all theuser sets are clustered according to the microblog user characteristics to acquire potential influence user sets in the microblog; (3) a tensor influence model under the specific topic area is constructed for the potential influence user sets; (4) a new tensor influence model is reconstructed through utilizing tensor decomposition; (5) a user influence score is predicted based on the reconstructed tensor influence model; and (6), according to the user influence score, ranking from high to low scores is carried out, and the user influence ranking result is acquired and outputted. The method isadvantaged in that user influence prediction precision under the specific topic area can be improved. 本发明涉及种融合用户时间信息的微博客用户影响力排名方法，包括步骤：（1）对微博客用户进行特征化处理；（2）根据微博客用户特征对所有用户集进行聚类，得到微博客中的潜在有影响力用户集；（3）对潜在有影响力用户集构建特定话题领域下的张量影响力模型；（4