陈晓明, 李玲俐, 梁雄友. 消除语义网络中分词歧义方法研究[J]. 微电子学与计算机, 2012, 29(3): 178-181.
引用本文: 陈晓明, 李玲俐, 梁雄友. 消除语义网络中分词歧义方法研究[J]. 微电子学与计算机, 2012, 29(3): 178-181.
CHEN Xiao-ming, LI Ling-li, LIANG Xiong-you. Eliminate Semantic Network Word Segmentation Ambiguity Method Research[J]. Microelectronics & Computer, 2012, 29(3): 178-181.
Citation: CHEN Xiao-ming, LI Ling-li, LIANG Xiong-you. Eliminate Semantic Network Word Segmentation Ambiguity Method Research[J]. Microelectronics & Computer, 2012, 29(3): 178-181.

消除语义网络中分词歧义方法研究

Eliminate Semantic Network Word Segmentation Ambiguity Method Research

  • 摘要: 针对计算机语义网络中交集型和组合型词汇岐义的问题, 通过分析了传统分词方法中存在的缺陷, 提出基于最大概率计算的自动分词歧义方法.运用上下文语义相关度对产生歧义的词汇进行有效修正, 重新计算切分候选词所产生的有效“费用”, 运用最大概率计算法对产生歧义的词汇进行关联程度概率计算, 克服传统分词方法的弊端.成功地解决交集型岐义、连环交集型岐义、组合型岐义、混合型岐义切分问题, 消除语义网络中的交集型和组合型词汇岐义的影响, 取得了不错的效果.

     

    Abstract: According to the computer in the semantic networks combination of overlap type out the problem of righteousness vocabulary, through the analysis of the traditional word segmentation method defects.Based on the maximum probability calculation method of automatic word segmentation ambiguity.Application for the semantic relatedness produces different meanings to the context of vocabulary, effective correction to the calculation of the segmentation candidates from the word of effective "expenses", the most probability calculation method to produce ambiguity of vocabulary connection degree probability calculation, overcome the shortcomings of the traditional word segmentation method.Successfully solve overlap type righteousness, serial overlap type, combination and mixed segmentation problem, eliminate righteousness in the semantic networks and combination of overlap type out the effect of vocabulary righteousness, and achieved good effect.

     

/

返回文章
返回