基于RFE+CatBoost模型的异常用电检测方法研究

Research on Abnormal Electricity Detection Method Based on RFE+CatBoost Model

针对传统电力检测领域中异常用电检测模型需要调节大量超参数导致异常用电检测效率低下,以及模型选取特征不能充分反映实际用电情况导致分类精度不高的问题,提出一种基于RFE+CatBoost模型的异常用电检测方法。较传统的异常用电检测方法而言,CatBoost算法降低了模型检测对于超参数的依赖。以用户用电数据作为研究对象,结合RFE算法分析用户在用电表现上的不同特征,采用分类预测算法对异常用电行为进行进一步研究,最后通过云南某地用户用电数据集进行验证,与其他用电异常检测模型进行对比,实验证明所提模型具有很好的检测能力,对于提升企业用电异常检测效率、指导用户更好地用电具有重要意义。

Aiming at the problem that the abnormal power usage detection model in the traditional power detection field needs to adjust a large number of hyperparameters,which leads to low efficiency of abnormal power usage detection,and the classification accuracy is not high due to the model selection characteristics that do not match the actual power usage.A method for detecting abnormal electricity consumption based on the RFE+CatBoost model is proposed.Compared with the traditional abnormal electricity usage detection method,the CatBoost algorithm reduces the dependence of model detection on hyperparameters.Taking user electricity data as the research object,analyze the different characteristics of users in terms of electricity performance combined with the RFE algorithm.Finally,it was verified by the electricity data set of a certain place in Yunnan,Compare with other anomaly detection algorithms,Experiments show that the proposed model has a good ability to detect abnormal electricity consumption,it is of great significance to improve the efficiency of electricity anomaly detection and guide users to better use electricity.