作 者: ;
机构地区: 广东医科大学
出 处: 《情报杂志》 2014年第1期157-160,共4页
摘 要: 在深入分析医学主题词标引规则基础上,以副主题词的组配作为主题词词性识别的依据,以副主题词的组合作为主题词之间语义表达的方向,构建“药物”治疗“疾病”的语义结构模板。并以肿瘤学语料库为样本,通过词频分析法与多维定位方法,批量识别出学科中“药物”治疗“疾病”的词汇语义关系。通过实例验证,两种方法均能实现这种类型的词汇语义关系识别。但多维定位法,具有方法简单,准确率和召回率高的特点,是一种值得推广的方法。 This paper constructs a semantic template of drug-therapy-disease based on the in-depth analysis of the Medical Subject Head-ings Indexing rules, and identifies semantics relations in batch of“drug” curing“disease” by word frequency analysis and multidimension-al positioning method in oncology corpus sample. Verified by an example, both of the two methods can achieve this type of lexical seman-tic relations identification. However, multidimensional positioning method, with the characteristics of simplicity, high accuracy rate and high recalling rate, is a worthy approach.
分 类 号: [G254]
领 域: []