本文已被：浏览 1259次 下载 550次
投稿时间：2013-09-13
投稿时间：2013-09-13
中文摘要: 异常检测问题是不均衡分类问题,Adaboost算法是一种有效的分类方法.分析了标准Adaboost算法,找出了标准Adaboost算法两个可以改进的地方,给出了改进的Adaboost算法,并在此基础上给出了异常检测算法.对医疗数据的异常检测结果表明了该算法的有效性.
中文关键词: 异常检测 Adaboost算法 减少抽样 聚类
Abstract:Outlier detection is an imbalance classification issue,while the adaboost algorithm is an effective classification method.The process of the original adaboost algorithm is analyzed and it is found that two methods can be improved.Then the improved adaboost algorithm is presented and based on this,the outlier detection algorithm is put forward.Finally,experiments are performed on widely used datasets WDBC and the result shows our algorithm is effective.
文章编号：20130611 中图分类号： 文献标志码：
基金项目:
引用文本：
张安勤,叶文琚.基于改进的Adaboost算法的异常检测[J].上海电力大学学报,2013,29(6):558-562.
ZHANG Anqin,YE Wenjun.Outlier Detection Based on the Improved Adaboost[J].Journal of Shanghai University of Electric Power,2013,29(6):558-562.
张安勤,叶文琚.基于改进的Adaboost算法的异常检测[J].上海电力大学学报,2013,29(6):558-562.
ZHANG Anqin,YE Wenjun.Outlier Detection Based on the Improved Adaboost[J].Journal of Shanghai University of Electric Power,2013,29(6):558-562.