基于维基百科的军事舆情论坛话题追踪方法

摘要

针对互联网论坛话题追踪,提出一种基于维基百科知识的军事话题追踪方法.该方法首先以基于维基百科的词语语义相关度与共现统计方式,同时结合军事主题与帖子的结构特征建立文本图中节点间的关系边及其权重；接着以改进的基于图的链接挖掘方法选取帖子关键词；最后通过计算话题与文本关键词列表间的语义相关度实现话题追踪.实验表明,该方法无需大规模样本训练与语义知识的手工构建,能够有效解决语义稀疏对追踪所带来的负面影响,较好地追踪到军事话题帖.%A method using Wikipedia as semantic and background knowledge was proposed for public military topic tracking on BBS. The semantic profiles of a post was modeled by text graph, in which nodes and edges were considered as: Wikipedia-based words semantic relevance, word co-occurrence with military themes and post structure, then a modified link mining method was utilized to extract the key words from text graph. At last, topic tracking was realized by calculating the semantic relevance of keywords between the post and topic. In the experiment, the results show that this method can effectively solve the problem of semantic feature scarcity in BBS-oriented military topic tracking.