常见新近决策树算法及其在卫生领域中的应用  

Common recent decision tree algorithms and the applications in the field of health

作　　者：林文怡 宛小燕[1] 刘元元[1] LIN Wen-yi;WAN Xiao-yan;LIU Yuan-yuan(West China school of Public Health and West China Fourth Hospital,Sichuan University,Chengdu,Sichuan 610041,China)

机构地区：[1]四川大学华西公共卫生学院/四川大学华西第四医院

出　　处：《现代预防医学》2019年第23期4233-4237,4242,共6页Modern Preventive Medicine

基　　金：四川省科学技术厅项目(编号:2019YJ0076)

摘　　要：目的综述常见新近决策树算法及其在卫生领域的应用进展,为后续相关研究提供参考。方法以"数据挖掘"、"决策树算法"、"decision trees"等作为检索词,检索2010-2019年CNKI、万方、维普、PubMed等数据库,总结随机森林、C5.0、GBDT等方法的基本原理、步骤、适用条件及优缺点,列举其在卫生领域中的应用。结果不同决策树算法各有其优势,在卫生领域多学科中有不同程度的应用。结论新近决策树算法较原始算法有较大改进,但仍存在不足。推广新近算法,针对其缺陷进行改进,将是决策树算法未来研究的重要方向之一。Objective To summarize the common recent decision tree algorithms and their applications in the field of health, and to provide reference for subsequent researches. Methods "Data mining", "decision trees" were used as searching terms in CNKI, Wanfang, Weipu, PubMed database and so on from 2010 to 2019. The basic theory, steps, applicable conditions, advantages and disadvantages, applications in the field of heath of random forests, C5.0, GBDT and so on were summarized and listed. Results Different decision tree algorithms had their own advantages, and they had different degrees of applications in many subjects of health. Conclusion The recent decision tree algorithms are much better than the original algorithms, but they still have some defects. Promoting the applications of recent algorithms and improving their defects will be one of the important directions in the future.

关 键 词：数据挖掘 决策树 新近算法 

分 类 号：R195[医药卫生—卫生统计学]