摘要: 在话题追踪研究领域,话题随着时间不断发展变化。目前的话题追踪方法无法对话题的发展演化进行全局的把握。针对该问题,提出基于相似度计算的话题演化分析方法。该方法采用时间片划分的思想,通过子话题间的相似度计算得到话题演化的具体过程及细节。实验结果表明,该方法能有效地反映话题的演化历程。
关键词:
话题追踪,
时间片,
子话题,
话题演化,
文本相似度
Abstract: In the area of topic tracking, topic develops with time, traditional topic tracking method can tracking the relevance story efficiently. It can not know relationship between events occurred during the develop of topic. It can neither know the whole history of the topic tracking. This paper proposes a new method based on the calculation of the subtopic similarity. The method concern the time characteristic of the topic tracking and use it to manipulate the information. Experiment result shows that the method analysis the evolution history of the topic efficiently.
Key words:
topic tracking,
time slice,
subtopic,
topic evolution,
text similarity
中图分类号:
吕 楠;罗军勇;刘 尧;杨慧洁. 基于话题三层结构模型的话题演化分析算法[J]. 计算机工程, 2009, 35(23): 71-72,7.
LV Nan; LUO Jun-yong; LIU Yao; YANG Hui-jie. Topic Three Layer Model Based Topic Evolution Analysis Algorithm[J]. Computer Engineering, 2009, 35(23): 71-72,7.