以阿司匹林副作用为研究样本的主题,搜集PubMed数据库中与“药物副作用引起疾病”主题词关联规则相对应的文献记录,在文献摘要中提取出相应实体共现的句子,进而利用自然语言处理方法提取文献中表达生物医学实体间语义关系的动词,共提取出30个表示药物副作用与疾病关系的动词。研究结果表明:基于共现和自然语言处理提取语义关系动词的方法是可行的。 Taking the aspirin side effects as the research topic, we collected the records of PubMed database corresponding to the keywords association rules of “disease caused by side effects”, extracted the co-occurrence sentences of the corresponding entities in the literature summary, and then used the natural language processing Methods Verbs used in the literature to express the semantic relationships between biomedical entities were extracted and 30 verbs were selected to indicate the relationship between drug side effects and diseases. The results show that it is feasible to extract semantic relational verbs based on co-occurrence and natural language processing.