摘要： 基于编辑距离和多种后处理的生物医学文献实体名识别方法通过“全称缩写对识别算法”扩充词典，利用编辑距离算法提高识别召回率。在后处理阶段，使用前后缀词扩展、POS扩展、合并邻近实体及利用上下文线索等方法进一步提高性能。实验结果表明，使用该方法即使利用内部词典也可以获得较好的识别效果。

关键词: 文本挖掘, 实体识别, 编辑距离, 条件随机域

Abstract: A bio-entity name recognition approach using edit distance and multiple postprocessing methods is presented, which expands dictionary via the abbreviation definitions identifying algorithm and improves the recall rate through the edit distance algorithm. The post-processing methods of improving the performance including first-keywords and post-keywords expansion, POS(Part of Speech) expansion, merge of adjacent entity names and the exploitation of the context cues are discussed. Experimental results show that with the above methods even an internal dictionary based system can achieve a fairly good performance.

Key words: text mining, entity recognition, edit distance, Conditional Random Fields(CRF)

中图分类号: 

• TP391.4