基于条件随机场的中医术语抽取方法及其应用探析

中医文献有种类繁多,数量庞大,记录随意,术语表达方式独特等的特点,为知识的获取带来困难。信息抽取技术可以利用计算机对文本信息进行针对性抽取,以结构化的形式将结果储存到数据库中,这种技术可以帮助医学研究者从海量信息中高效获取所需知识。命名实体识别是信息抽取准确与否的关键,对目前常用的几种识别方法进行分析,认为基于统计的方法更适用于中医文献的研究,并选定条件随机场算法,结合中医术语的特点,对该方法及步骤进行了详细阐述。同时,举例介绍了信息抽取技术在中医结构化电子病历及中医专业领域搜索引擎建立中的辅助作用,为其在中医领域的应用提供更广阔的参考思路。

TCM Terminology Extraction Method and Application Based on Conditional Random Field

Literature of traditional Chinese medicine has a record of variety,large quantity,arbitrary,unique professional term expression,and so on. These characteristics brought difficulties for knowledge acquisition. Information extraction technology can make use of the computerto give you text information extraction,and then the results in the form of structured stored in the database. This technique can help medical researchers from large amounts of information quickly to get the knowledge they need. Named entity recognition is the first step in information extraction technology. In this paper,analysis of the current commonly- used several kinds of recognition methods,thinking statistical method is suitable for the research of TCM literature and decision in the conditional random field method,combining with the characteristics of TCM terms,to this kind of method in detail in this paper. In addition,it introduces two applications of information extraction technology in the traditional Chinese medicine：the TCM structured electronic medical records and the TCM professional search engine,helping the field of information extraction technology in traditional Chinese medicine to provide a broader application range.