Study On The Mapping Of Colloquial Names Of Diseases To Terminology Names

[Purpose]With the further development of artificial intelligence and "Internet + medical",there is a need to deal with colloquial names of diseases in platforms or applications such as online consultation,intelligent consultation and electronic medical records.At present,there are few mapping analyses of colloquial names and terminology names of diseases in China.This paper aims to enrich the mapping research of colloquial names and terminology of patients’ diseases,and to solve the problem that it is difficult for platforms and tools such as online consultation and intelligent guidance to make effective use of colloquial names of online diseases.[Methods]To grab the colloquial disease names from the comments of patients on Guahao.com,we will use Pubmed Chinese medical literature text,Baidu search page abstract and Baidu encyclopedia content as corpus.We will train the word embedding model,use the cosine distance to calculate the similarity of the colloquial names and terminology names of diseases,and adjust the value of the parameter TopN to determine the optimal mapping result.Throughout the experiment,mapping will be done with the help of an intermediate vocabulary and without the help of an intermediate vocabulary,and each mapping includes the mapping of the ICD-11’s main category and subcategory classification terminology names.The accuracy of the results will be tested by two coders with clinical knowledge and professional experience.[Results]When the parameter TopN of the most similar entry is set to N=30,the effect is better.The accuracy of mapping to the main category is stable at about 60%,and the accuracy of mapping to the subcategory classification is about 70%.After using the intermediate word list for indirect mapping,the experimental results show that the accuracy rate of the major category and subcategory mapping is more than 80%.[Conclusions]The findings of this paper show that the use of word embedding to map colloquial names of disease and disease terminology names has strong feasibility.The intervention of the intermediate word list is of great significance for the mapping of the international disease classification ICD-11.The paper has reference value for online consultation,intelligent guidance and automatic coding in electronic medical records.Meanwhile,a more comprehensive and accurate corpus and more detailed calculation rules are conducive to improving the accuracy of related research.