非结构化病理文本的结构化信息抽取方法

Method of Structured Information Extraction from Unstructured Pathological Texts

介绍病理文本数据结构和概念层次结构,以非结构化的病理文本为对象,首先对非序病理文本的结构进行分析,其次利用模式匹配对病理文本予以模式提取和泛化,最后从分词序列中抽取结构化信息,实验表明该方法能够获得较高的准确率和召回率。

The paper introduces the data structure and conceptual hierarchy of pathological texts.Based on unstructured pathological texts,at first,it analyzes the structure of pathological texts other than prefaces.Next,it conducts pattern extraction and generalization of the pathological texts by pattern matching.At last,it extracts structured information from the participle sequence.As proved by the experiment,this method can achieve a high accuracy and recall rate.