摘要
针对医疗体系中对患者违约行为约束力弱、优质医疗资源浪费严重的问题。文中基于机器学习算法对失信医疗信息防御和监测机制进行了研究,设计了基于随机森林(RF)算法的失信识别模型。文中设计的随机森林算法基于CART树,使用基尼系数作为节点的判别分类标准,提升了传统决策树的收敛速度。由于随机森林算法基于集成学习的Bagging思想,可以有效避免训练过程的过拟合现象,提升了单一决策树算法的分类精度。数据测试结果表明,相较于逻辑回归、K-邻近等其他机器学习算法,该模型在分类精度上分别可以提升1.3%和1.4%,可以为社会医疗信用体系的建立与完善提供技术支撑。
In view of the weak binding force on patients'breach of contract and the serious waste of high-quality medical resources in the medical system.In this paper,based on machine learning algorithm,the mechanism of information defense and monitoring of dishonest medical treatment is studied,and a model of dishonest recognition based on random forest(RF)algorithm is designed.The random forest algorithm designed in this paper is based on cart tree and uses Gini coefficient as the classification standard of node discrimination,which improves the convergence speed of traditional decision tree.In addition,because the stochastic forest algorithm is based on the bagging idea of integrated learning,it can effectively avoid the over fitting phenomenon in the training process and improve the classification accuracy of single decision tree algorithm.The data test results show that compared with other machine learning algorithms such as logistic regression and K-proximity,the model can improve the classification accuracy by 1.3%and 1.4%respectively,which can provide technical support for the establishment and improvement of social medical credit system.
作者
高晓娟
GAO Xiao-juan(Information Department,Children’s Hospital of Nanjing Medical University,Nanjing 210008,China)
出处
《电子设计工程》
2020年第17期1-5,共5页
Electronic Design Engineering
基金
国家自然科学基金(61372071)。