Research on Semantic Extraction Method in Chinese Electronic Medical Records

In China, the technology of information extraction is widely used in many fields and the most notable one is medical science. Chinese electronic medical records(CEMRs) contain tremendous information and is of great value, however, free text like CEMRs cannot be directly used for analysis. In order to extract valuable information from CEMRs, this paper put forward a method to extract key information from unstructured CEMRs and has achieved preferable results. This paper is composed of three parts. The first part proposes semantic extraction model which is composed of three parts, namely, text segmentation, semantic disambiguation and semantic extraction. The following part explains the three modules of this model in detail. In order to improve semantic extraction accuracy, this paper proposes the algorithm of semantic disambiguation which is meant for CEMRs. Equipped with the sentence pattern library, Semantic extraction module is designed to extract key information from CEMRs and populate templates with extracted information in order to generate XML output. The last part is about experiment. The method proposed in this paper can be validated through experiment and result of the experiment can be applied to evaluate this method.