医疗健康知识挖掘中的语义资源、数据集和工具-《计算机技术与发展》

文章信息/Info

Title:: Semantic Resource,Dataset and Tool for Medical Health Knowledge Mining

Author(s):: ZHANG Wei; ZHANG Zhan-peng; ZHANG Ming-tao; HAN Pu; School of Management,Nanjing University of Posts & Telecommunications,Nanjing 210003,China

Keywords:: knowledge mining; semantic resource; dataset; data annotation; deep learning

摘要:: 医疗健康知识挖掘在人工智能和大数据时代受到了学界的极大关注, 目前已经成为信息抽取和文本挖掘中的重要研究方向。在基于深度学习的实体识别、实体关系抽取、问答系统以及知识图谱构建研究中,各类语义资源、数据集和工具已经成为开展医疗健康知识挖掘的重要保障。首先对医疗健康知识挖掘中需要使用的 UMLS、MeSH 和SNOMED CT 等语义资源进行了系统梳理,并详细分析了各类语义资源的实际应用场景,指出了中文语义资源存在的问题和不足;其次对英文和中文的电子病历、医学文献和在线健康数据集进行了重点论述,并对数据集的应用任务进行了分析;最后论述了常见的医疗健康文本处理工具和系统,并就其具体应用进行了讨论。为国内更好地开展医疗健康知识挖掘提供了参考。

Abstract:: Medical and health knowledge mining has received great attention from the academic community in the era of artificial intelligence and big data,and has now become an important research direction in information extraction and text mining. In the research of entity recognition,? entity relationship extraction,question answering system and knowledge graph construction based on deep learning,various semantic resources, datasets and tools have become an important guarantee for the development of medical and health knowledge mining. We firstly systematically sort out semantic resources such as UMLS,MeSH and SNOMED CT that used in medical and health knowledge mining,and analyze in detail the actual application scenarios of various semantic resources,and point out the problems and shortcomings of Chinese semantic resources. The electronic medical records,medical literature and online health data sets in English and Chinese are emphasized, and the application tasks of the data sets are analyzed. Finally,common medical and health text processing tools and systems are discussed, and their specific applications are discussed. It provides a reference for the better development of medical and health knowledge mining in China.