桂林电子科技大学学报2017，Vol.37Issue(3)：208-211,4.
一种鲁棒的局部异常检测算法
A novel robust local outlier detection algorithm
摘要
Abstract
The anomaly detection algorithm is very sensitive to the parameter selection,a robust local outlier detection algorithm is proposed.This novel method not only can effectively identify global and local outliers but also is associated with one statistical parameter.Furthermore,the sole parameter can be easily determined without relying on users'' domain knowledge.The method is insensitive to parameter variation,which ensures its superiority in robustness.In addition,the new algorithm is used to detect outliers of user electrical behavior,which can help power enterprises determine whether the user exists electricity stealing behavior.关键词
异常检测/数据挖掘/鲁棒/用电行为Key words
outlier detection/data mining/robust/electrical behavior分类
信息技术与安全科学引用本文复制引用
康欢,张爱梅,雷景生,宋硕..一种鲁棒的局部异常检测算法[J].桂林电子科技大学学报,2017,37(3):208-211,4.基金项目
国家自然科学基金(61472236,61672337) （61472236,61672337）
