摘要
为直接高效地获取文献中的知识,命名实体识别用来识别文本中具有特定意义的实体。这是应用文本挖掘技术自动获取知识的关键的第一步,因此受到日益广泛的关注。主要从评测方法、特征选择、机器学习方法和后期处理等方面介绍了近年来生物医学命名实体识别方面的主要研究方法及成果,并对目前各方面存在的问题进行了分析和讨论,最后对该领域的研究前景进行了展望。
In order to acquire the useful knowledge directly and effectively from documents, named entity recognition is to recognize the meaningful entities in documents. It is the first and important step to acquire relevant knowledge automatically by using text mining technology. This review introduced main approaches and achievements in recognizing biomedical named entities, focusing on the strategy of evaluation, feature selection, methods on machine learning and the post-processing. It also looked into the current problems and displayed the promising solutions. The last section drew the prospection for the research on biomedical named entity recognition.
出处
《计算机应用研究》
CSCD
北大核心
2010年第3期811-815,832,共6页
Application Research of Computers
基金
国家自然科学基金资助项目(60773021
60603054)
关键词
命名实体识别
文本挖掘
特征选择
机器学习
named entity recognition
text mining
feature selection
machine learning