计算机与数字工程2017，Vol.45Issue(12)：2398-2401,2414,5.DOI:10.3969/j.issn.1672-9722.2017.12.016
基于多特征融合的窃电识别算法研究
Study on Stealing Recognition Model Based on Multi Feature Fusion
邓明斌 1徐志淼 1邓志飞 1张卓 1迟艳杰1
作者信息
摘要
Abstract
For the present situation,the stealing power methods of specialized users are various.At the same time,the feasi?bility of traditional anti-stealing power measures is low. A scheme of abnormal electricity intelligent analysis model based on multi-feature fusion is provided for this phenomenon. This scheme extracts the features of abnormal power performance and then judge the feature with threshold which is trained by machine learning methods.By fusing the judged result of each feature with the multi-feature fusion,the result can be transformed to abnormal exponent for screening of abnormal based on abnormal threshold.By this method,abnormal electricity users will be screened out basically.关键词
用电行为/特征提取/多特征融合/窃电识别Key words
electrical behavior/feature extraction/feature fusion/stealing recognition分类
信息技术与安全科学引用本文复制引用
邓明斌,徐志淼,邓志飞,张卓,迟艳杰..基于多特征融合的窃电识别算法研究[J].计算机与数字工程,2017,45(12):2398-2401,2414,5.
