【摘要】 电子病历EMR(Electronic Medical Records)检索是信息检索研究中的一个新领域。医学术语在电子病历检索中占有重要地位,通常用来限定检索条件、表达用户的检索意图。针对这种情况,提出一种基于医学术语权重调整的查询重构方法,以提高电子病历检索的性能。该方法首先从原始查询语句中筛选出医学术语,然后使用自信息来度量每个医学术语的权重,最后将加权的医学术语与原始查询语句按照一定的权重比例结合,构造出新的查询语句。将该方法在TREC数据集上进行实验,结果表明与原始查询结果相比,重构后的查询结果在MAP、bpref和P10这三项指标上,分别提高了14.2%、10.1%和9.6%,验证了该方法的有效性。更多还原

【Abstract】 Electronic medical record( EMR) retrieval is a new field in information retrieval. Medical terms occupy an important position in EMR retrieval,and they are usually used to limit the retrieval conditions and to suggest users’ search intention. Aiming at the importance of medical terms,we proposed a medical terms reweighting-based query reformulation method to improve the performance of EMR retrieval.First,the method screens out medical terms from original query sentences,and then measures the weight of each medical term by its own selfinformation. Finally, it constructs new query sentences by combining the weighted medical terms and the original query sentences proportionally. Our method has been experimented on TREC dataset, results showed that compared with original query results, the reformulated query results improved in three items of MAP( + 14. 2%),bpref( + 10. 1%) and P10( + 9. 6%) respectively,and this verified the effectiveness of the method.更多还原