Author Login Editor-in-Chief Peer Review Editor Work Office Work

Computer Engineering ›› 2011, Vol. 37 ›› Issue (24): 45-49. doi: 10.3969/j.issn.1000-3428.2011.24.015

• Networks and Communications • Previous Articles     Next Articles

Topic Thread Detection Algorithm in Dynamic Text Flow

CAO Yue-qin   

  1. (Department of Computer Science, Wenzhou Vocation & Technical College, Wenzhou 325027, China)
  • Received:2011-09-02 Online:2011-12-20 Published:2011-12-20

动态文本流中的话题线索检测算法

曹月芹   

  1. (温州职业技术学院计算机系,浙江 温州 325027)
  • 作者简介:曹月芹(1971-),女,硕士,主研方向:数据挖掘,模式识别
  • 基金资助:
    浙江省教育厅基金资助项目(Y200909362)

Abstract: For the topic clue inspection of dynamic text flow, this paper proposes an inspection algorithm based on semantic relevance model. It calculates the semantic relevance of message text and topic clue, allots message to related topic clue, and adopts the improved algorithm of semantic relevance model to inspect different data sets and topics of different lengths. Experimental results show that this algorithm gives higher F1 value than current algorithms.

Key words: dynamic text flow, topic thread detection, semantic relevance degree, message division, Instant Messaging(IM) tool

摘要: 针对动态文本流中的话题线索检测问题,提出一种基于语义相关度模型的话题线索检测算法。利用改进的语义相关度模型计算消息文本与话题线索之间的语义相关度,将消息划分至相应的话题线索。在真实数据集上的实验结果表明,与现有算法相比,该算法的F1测度值较高。

关键词: 动态文本流, 话题线索检测, 语义相关度, 消息划分, 即时通信工具

CLC Number: