Publishing review comments of micro-blog hot topic is becoming the daily round in modern smart city. With the in-depth discussion about the micro-blog topic, it is a challenge for users to understand quickly the basic content of the topic due to its gradually complex inner relationship. To solve the problem, this paper presents the construction of micro-blog topic summarization. First, the keywords are extracted based on the lexical features of words occurred in comments published by users. And then the association rules between keywords are extracted by Apriori. Second, the associated semantic network for micro-blog topic (ASN-MT), the visualization of micro-blog topic summarization, is constructed according to the extracted keywords and association rules. Third, ASN-MT is optimized by reducing its scale based on the confidence threshold. And then its optimal value is selected according to the integrity of ASN-MT. The experimental results show that the proposed algorithm can construct accurately and quickly a concise and complete micro-blog topic summarization.

微博热点话题发表评论正在成为现代智慧城市的日常活动。随着对微博客主题的深入讨论，由于其内部关系逐渐复杂，用户快速理解主题的基本内容是一个挑战。为了解决这个问题，本文提出了微博主题摘要的构建。首先，根据用户发表的评论中出现的单词的词汇特征提取关键词。然后由Apriori提取关键词之间的关联规则。其次，根据提取的关键词和关联规则，构建了微博主题的语义网络（ASN-MT），即微博主题摘要的可视化。第三，通过基于置信度阈值减小其规模来优化ASN-MT。然后根据ASN-MT的完整性选择其最佳值。实验结果表明，该算法可以准确，快速地建立简洁，完整的微博主题摘要。