基于PCA孤立森林的用电异常识别研究 |
点此下载全文 |
引用本文:何 昆,何子昂,范杏元.基于PCA孤立森林的用电异常识别研究[J].计算技术与自动化,2021,(2):76-80 |
摘要点击次数: 445 |
全文下载次数: 0 |
|
|
中文摘要:用电异常是供电管理工作中一大痛点,传统的工作管理模式,对于用电异常的发现存在偶然性,无法保证工作效率。本文响应智能电网建设的号召,充分利用数据挖掘手段进行用电异常智能识别研究。在真实数据的支撑下,经反复尝试、推演,得出所提出的基于PCA的孤立森林用电异常识别方法。并且,在孤立森林、PCA-逻辑回归、PCA-神经网络的对比实验验证下,所述方法依然具备优越性和进步性。 |
中文关键词:用电异常 时间序列数据 PCA孤立森林 仿真实验 |
|
Research on Electricity Abnormal Identification of Isolated Forest Based on PCA |
|
|
Abstract:Power abnormality is a major pain point in power supply management. The traditional work management mode has contingency for the discovery of power anomalies and cannot guarantee work efficiency.This paper conforms to the purpose and trend of smart grid construction, and makes full use of data mining methods to conduct intelligent anomaly intelligent identification research. Under the support of real data, after repeated trials and deductions, the PCA-based isolated forest electricity anomaly identification method is derived.In addition, under the comparative experiment of isolated forest, PCA-logical regression and PCA-neural network, the method described still has superiority and progress. |
keywords:abnormal power corsumption time series data PCA isolation forest simulation experiment |
查看全文 查看/发表评论 下载pdf阅读器 |