学术论文

      基于双模式的产品指标本体概念抽取

      Ontology concept extraction for product indicators based on double strategy combination

      摘要:
      由于人们对产品指标执行情况的关心日益增加,人们需要一种能够清晰描述产品指标的方法,而以本体为基础描述产品指标的方法由于其清晰的领域知识结构能够为统计分析打好基础成为目前研究者研究的主要方向之一.为了减少产品指标本体构建对领域专家的依赖度,文中提出了一种双策略结合的产品指标本体概念抽取模式,针对产品指标数据的结构特点,融合两种分类特征重叠范围较小的策略分别对关键词赋予权重并加权排序,之后可由人工挑选或设定阈值确定领域概念.实验证明使用该方法抽取概念相较于单一策略准确性有所提高,且抽取结果是有效的.
      Abstract:
      Due to the increasing concerns about the implementation of product indicators,people need a method which can clearly describe the product indicators.The method of ontology can clear the knowledge structure of domain and build the foundation for statistical analysis.So the ontology-based method has become one of the main research directions of the researchers.To reduce the dependence of ontology construction on domain experts,a method of concepts automatic extraction based on two strategy combination is proposed.The two strategy that this method choosed have a little overlapping rangeare of characteristcs,and aimed at the structural characteristics of product index data.This method uses two strategies to calculate the weight of candidate words and ordering them,then could set threshold or select manually to identify domain concepts.After lots of experiments,the feasibility and validity of this method are proved.
      作者: 时公泽 王浩畅
      Author: SHI Gong-ze WANG Hao-chang
      作者单位: 东北石油大学计算机与信息技术学院,黑龙江大庆,163000
      刊 名: 信息技术 ISTIC
      年,卷(期): 2017, (3)
      分类号: TP391
      在线出版日期: 2017年4月18日
      基金项目: 国家自然科学基金