汉语名词语义特征建模与分析

发布时间：2019-02-14 10:22

【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]:Objective: to construct a semantic feature model of Chinese nouns to provide a quantitative and visual semantic feature database for clinical speech therapy. Methods: using 60 common noun concepts as stimuli, the healthy subjects were nominated for semantic features, and the semantic feature items of each concept were collected, and then the items were classified according to the Chinese semantic feature classification scheme. Finally, the classification data visualization, clustering and statistical test are implemented with R software. Results: 1Conceptual clustering based on semantic features is basically consistent with empirical classification results. Semantic features can reflect semantic relationships between concepts. 2 semantic cognitive processing in different conceptual domains has type bias. Biological concepts have more sensory information features, while non-biological concepts have more functional features. Conclusion: the model based on Chinese semantic feature data can effectively reflect the conceptual semantic structure and is helpful to extract semantic training materials according to quantitative indexes.
【作者单位】：南京医科大学第一附属医院康复医学科;南京师范大学语言科技研究所江苏省哲学社会科学重点研究基地;
【基金】：江苏省科技支撑计划(BE2012675) 国家自然科学基金资助项目(81171854)
【分类号】：H136;R493

【参考文献】