离群点分析在高校能耗监控系统中的应用

节能是当今社会面临的重大课题,高校作为能源大户以及教书育人的基地,必须在能耗监控系统中起到领先示范的作用。在节能分析系统中,能耗的预警预测是关键,因此,异常点的发现与分析,为预警提供了直接的依据,是整个系统的基础。在数据挖掘中,离群点检测分析可以通过多种方法实现,本文应用了基于统计分布的离群点检测方法,但由于在实际情况中,能耗数据的变化与社会各类群体的生活习性、工作周期相关,这些复杂性决定了在数据分析中,只能根据实际的业务来检验分析结果的正确性。本文通过对某高校的能耗进行基于统计分布的离群点分析,并结合校园能耗规律,得出在高校中能耗的异常情况并报警,以达到节约能耗的目的。

Application of Outlier Detection in University of Energy Monitoring System

Energy conservation is a major issue in today＇s society,as the teaching base of energy-hungry,colleges must play a leading role in energy consumption monitoring system. Early warning and forecast energy consumption is the key in the energy analysis,thus detection and analysis outliers is the foundation of the whole system and provides a direct basis for early warning. In data mining,outlier detection has several ways to achieve. This paper is based on the statistical distribution of outlier detection methods.However,as in reality,changes of energy consumption data related with the living habits of social groups and the working cycle,the correctness of these complexities determined that it can only be based on actual business results of tests and analysis in data analysis.This paper finds the abnormal situation of the universities＇ energy consumption and then activates an alarm by the statistical distribution of outlier detection methods of the energyconsumption combined with consumption pattern in the university,in order to achieve the purpose of saving energy.