基于机器学习的针灸相关疾病、基因、药物新关联挖掘

Machine learning-based mining of new associations in acupuncture-related diseases, genes and drugs

中文关键词: 针灸  疾病  基因  药物  SVM  关联网络  数据挖掘  文本挖掘

英文关键词: Acupuncture  Disease  Gene  Drug  SVM  Association network  Data mining  Text mining

中文摘要:

      目的：挖掘针灸相关疾病、基因和药物间的新关联。方法:提出一种基于SVM的机器学习算法，结合词典识别疾病、基因和药物实体并挖掘三者之间的关联，构建针灸相关疾病、基因和药物关联网络。结果:识别出针灸相关的296种疾病、51种基因和278种药物，并在27种疾病、13种基因和135种药物之间挖掘出704种关联，构建3个关联网络，发现了262种新关联。结论:针灸相关疾病-基因-药物之间存在大量程度不一的关联，为针灸精准医疗提供了新的研究思路。

英文摘要:

      Objective To mine the new associations in acupuncture-related diseases, genes and drugs. Methods A support vector machine (SVM)-based machine learning algorithm was proposed and different association networks for acupuncture-related diseases, genes and drugs were established by identifying the diseases, genes and drugs with dictionaries and mining their associations. Results A total of 296 acupuncture-related diseases, 51 genes and 278 drugs were identified, and 704 associations were mined in 27 diseases, 13 genes and 135 drugs. Three association networks were established, which discovered a total of 262 new associations in acupuncture-related diseases, genes and drugs. Conclusion New associations are detected in acupuncture-related diseases, genes and drugs, which can thus provide certain new ideas for studying the precision treatment of diseases by acupuncture.