基于BP-AsymBoost的医疗诊断模型  

A medical diagnosis model based on BP-AsymBoost algorithm

作　　者：张涛[1,2] 郝晓玲[1,2] 张玥杰[3] 张明辉[1] 

机构地区：[1]上海财经大学信息管理与工程学院,上海200433 [2]上海市金融信息技术研究重点实验室,上海200433 [3]复旦大学计算机科学技术学院上海市智能信息处理重点实验室,上海200433

出　　处：《系统工程理论与实践》2017年第6期1654-1664,共11页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71171126;71401096;61572140);教育部高等学校博士学科点专项科研基金(20130078110001)~~

摘　　要：本文主要研究BP神经网络以及AdaBoost算法在医疗诊断中的应用,在分析标准AdaBoost算法的基础上提出改进的AdaBoost算法,即BP-AsymBoost.针对UCI数据库中的威斯康星乳腺癌数据集设计结合BP网络以及改进后的AdaBoost算法的诊断模型,并且通过多个指标将其与BP模型,遗传算法优化的BP模型,未经改进的BP-AdaBoost模型进行比较,验证BP-AsymBoost模型的有效性.Abstract This paper mainly researches the BP application of neural network and AdaBoost algorithmin medical diagnosis. A novel improved AdaBoost algorithm, BP-AsymBoost, is proposed based on the analysis of the standard AdaBoost algorithm. Based on the datasets from the Wisconsin breast cancer data in UCI database, the diagnosis model is designed combining with the AdaBoost algorithm and the improved BP network. Through the comparsion with the basic BP model, the BP model with genetic algorithm optimization, as well as the BP-AdaBoost model on a number of indicators, the validity of our improved BP-AdaBoost model can be effectively verifed. Keywords computer-aided-dignosis; classification model; AdaBoost algorithm; BP network; BP-AsymBoost model

关 键 词：智能辅助诊断 分类模型 ADABOOST算法 BP神经网络 BP—AsymBoost模型 

分 类 号：R441[医药卫生—诊断学]