首页> 外国专利> WORD LINKING IDENTIFICATION MODEL LEARNING DEVICE, WORD LINKING DETECTION DEVICE, METHOD AND PROGRAM

WORD LINKING IDENTIFICATION MODEL LEARNING DEVICE, WORD LINKING DETECTION DEVICE, METHOD AND PROGRAM

机译:单词链接识别模型学习装置,单词链接检测装置,方法和程序

摘要

To provide a word linking identification model learning device, a word linking detection device, a method and a program capable of accurately learning an identification model which can identify whether or not linking of words is natural.SOLUTION: A word linking identification model learning device 100 comprises: a seed extraction unit 30 which performs a morphological analysis on a text collection in a predetermined domain, extracts, as a seed, a word string obtained by using a predetermined threshold from the result of the morphological analysis and defines the seed as positive example date; a negative example data expansion unit 34 which performs morphological analysis on a replacement character string in which words included in the seed are replaced, specifies a portion where linking of words in the replacement character string does not match an original part-of-speech string from the result of the morphological analysis and generates negative example data; and an identification learning model 36 which, on the basis of the positive example data and the generated negative example data, learns a word linking identification model 40 for identifying whether or not linking of words in the word string is natural.SELECTED DRAWING: Figure 1
机译:为了提供单词链接识别模型学习设备,单词链接检测设备,方法和程序,能够准确地学习可以识别单词链接是否自然的识别模型。解决方案:单词链接识别模型学习设备100包括:种子提取单元30,其对预定域中的文本集合进行形态分析,从形态分析结果中提取使用预定阈值获得的单词串作为种子,并将其定义为肯定示例。日期;否定示例数据扩展单元34对替换了种子中包括的单词的替换字符串执行形态分析,该否定示例数据扩展单元34指定替换字符串中的单词链接与原始词性字符串不匹配的部分。形态分析的结果并产生阴性实例数据;识别学习模型36,其基于正例数据和所生成的负例数据,学习词链接识别模型40,该模型用于识别单词串中的词链接是否自然。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号