汉语名词语义特征建模与分析

关键词：名词概念  语义特征  聚类分析  秩次  康复

Keywords:feature norms  semantic feature  clustering analysis  order of production  rehabilitation

摘要：

      摘要 目的：构建汉语名词语义特征模型，为临床言语治疗提供量化和可视化语义特征数据库。 方法：以60个常见名词概念为刺激词，对健康受试者实施语义特征提名，采集各概念的语义特征条目，继而根据汉语语义特征分类方案对条目进行分类。最后以R软件实施分类数据可视化、聚类和统计检验。 结果：①根据语义特征进行的概念聚类与经验性分类结果基本一致，语义特征可以反映概念间语义关系。②不同概念领域的语义认知处理具有类型偏向性，生物概念具有更多的感觉信息特征，而非生物概念则具有更多的功能用途特征。③秩次居于首位的语义特征，在分类范畴型显著高。 结论：根据汉语语义特征数据建立的模型可以有效反映概念语义结构，有助于根据量化指标提取语义训练素材。

Abstract:

      Abstract Objective：To explore the model of semantic features for the Chinese nouns in order to provide quantitative and visual semantic database for clinical speech therapy. Method：The semantic feature entries were collected from a total of 60 stimulating nouns in volunteers by feature nomination. These entries were classified into different feature types according to Semantic Feature Classification Scheme for Chinese. With the R statistical computing environment, the distribution of feature types was visualized and a clustering analysis based on properties was conducted. Result：Firstly, the clusters based on properties were basically in line with the categories based on the experience. Secondly, cognitive processing of different domain had a certain bias. The living domain had significantly more perception properties，and nonliving domain had significantly more usage properties. Thirdly, the frequency of first rank features were significantly higher than expectation in the taxonomic categories. Conclusion：The model of semantic features for the Chinese nouns by feature nomination can effectively reflect the semantic structures of concepts and help to chose semantic training materials according to quantitative indicators.