口语化疾病名称向国际规范疾病术语集的映射研究

目的：研究患者口语化疾病名称到术语疾病名称的映射，解决在线口语化疾病名称难以有效利用的问题。方法：抓取“微医网”上患者口语中的疾病名，将PubMed摘要、百度搜索页摘要与百科内容作为语料，训练词向量，用余弦距离计算相似度，由2名具有临床知识和从业经验的编码人员对结果的准确性进行检验。结果：将最相似词条的参数TopN设置为>= 30的时候效果比较好，映射到大类和细类的准确性分别稳定在60%和70%左右。结论：利用词向量将口语化疾病名称映射到国际疾病术语集，具有较强的通用性和可行性，对于在线问诊、智能分诊甚至自动化编码具有明显的参考借鉴价值。

Mapping colloquial disease names into standard international glossary of disease terms

Objective To solve the difficulties in effective use of online colloquial disease names in patients by studying how to map the colloquial disease names into standard international glossary of disease terms. Methods The colloquial disease names used in patients were captured on "www.guahao.com". The word vectors were trained with the PubMed-covered abstracts, Baidu search page-covered abstracts and encyclopedia-covered contents used as the corpuses. Their similarity was calculated according to cosine distance with its accuracy tested by two experienced coding professionals. Results The mapping efficiency of colloquial disease names was good when at least 30 entries with the highest similarity were used. The accuracy of mapping colloquial disease names into the largest and smallest categories was about 60% and 70% respectively. Conclusion Mapping the colloquial disease names into standard international glossary of disease terms is of good universality and feasibility, and can thus play an important role in online consultation, intelligent diagnosis and automatic coding.