摘要
针对当前搜索引擎返回大量与用户请求无关结果的问题，提出一种新的生成相关搜索词的方法。该方法是基于开源搜索工具包Lucene的．通过对文档内容的切分和筛选找出概念词，再通过TF—IDF方法计算概念词的同现率，最后通过连续性Hopfield神经网络建立概念集，从而获得概念词的相关词推荐。
A new generation of term method is proposed for the current search engine problem that is return a large number of user requests has nothing to do with the outcome. The method is based on the Lucene open source search toolkit. First, it is find the term form segment and filter the document. Second, calculate the concurrence probability by the method of TF-IDF. Finally, set up the concept of set by the Hopfield neural network.
作者
吴佩韦
李昌华
WU Pei-wei, LI Chang-hua （School of Information and control, Xi＇an University of Architecture and Technology, Xi＇an 710055, China）
出处
《电脑知识与技术》
2009年第4期2612-2613,2624,共3页
Computer Knowledge and Technology
关键词
中文搜索引擎
相关词
概念词
同现率
概念集
chinese search engine
related words
the concept of the word
concurrence probability
the concept of set
