摘要
通过对医学文献中肺癌疾病与基因的关系进行分类研究可以掌握疾病特性,预测疾病变异,研发新型药品。针对Pub Med文献中肺癌疾病-基因关系分类的实际需求,首先进行关系抽取,其次分别对关系标志词进行特征表示与聚类,准确率达到91.60%,并与本体层对齐,验证了该方法的可行性与准确性。
By classifying the relationship between lung cancer diseases and genes in the medical literature,doctors can grasp the characteristics of disease,predict disease variation,and develop new drugs.This paper aims at the actual needs of lung cancer disease-gene relationship classification in Pub Med literature.First,the relationship extraction is performed.Secondly,the relationship markers are characterized and clustered separately.The accuracy rate is as high as 91.60%,and they are aligned with the ontology layer,which verified the feasibility and accuracy of the method.
作者
白萌
代晶
李鹏
BAI Meng;DAI Jing;LI Peng(不详;Information Department,General Hospital of Western War Zone,Chengdu 610083,Sichuan Province,P.R.C.)
出处
《中国数字医学》
2021年第10期103-107,共5页
China Digital Medicine
基金
四川省重点研发计划(编号:2021YFG0136).
关键词
肺癌
基因
关系抽取
关系分类
聚类
lung cancer
gene
relationship extraction
relationship classification
clustering
