Objective To achieve efficient matching and search of medical device classification and definition data,by designing a real-time interactive medical device classification recommendation system,based on the latest adjusted 2017 version of Medical Device Classification Catalogue.Methods The structured data from the 2017 version of Medical Device Classification Catalogue was extracted and associated database tables were created.Using a text similarity matching algorithm based on TF-IDF,the similarity between user input and example product names was calculated one by one to filter out the most matching content.At the same time,combining the association relationships of database tables,an interactive website was built using the Flask framework to display the medical device classification recommendation results to users.Results The classification recommendation system can automatically use the TF-IDF model to calculate text similarity based on the product name entered by the user,and return the highest matching key classification information such as the three-level classification recommendation code,subdirectory(main number,main name),first level product category(first level number,first level name),second level product category(second level number,second level name),product description,expected use,and management category through the webpage based on the association relationship of database table.Conclusion The medical device classification and recommendation system based on text similarity can provide convenient,accurate,and efficient guidance for medical device classification coding work,reduce classification confusion and failure caused by lack of professional knowledge and experience,and has good practical application value and guidance significance.
关键词
医疗器械/分类目录/文本相似度/分类推荐/网站开发
Key words
Medical devices/Classification catalog/Text similarity/Classification recommendation/Website development
维普
万方数据