Research On The Identification Of Abnormal Electricity Consumption Behavior Considering Data Imbalance

With the continuous construction of smart grids,smart meters can collect a large amount of basic electricity consumption data,which presents the characteristics of massive,high-dimensional,and imbalanced.Abnormal electricity consumption in electricity consumption data is defined as the amount of energy consumed abnormally much,reducing the country’s energy efficiency and adversely affecting a country’s economic situation.Identification of abnormal electricity consumption is essential to find abnormal electricity users,reduce economic losses,and effectively manage users’ electricity consumption.Therefore,it is necessary to carry out research on abnormal electricity consumption identification methods considering data imbalance.Considering the problem of abnormal electricity consumption identification of data imbalance,this paper takes abnormal electricity consumption data as an example of imbalanced data to carry out research on the identification method of abnormal electricity consumption behavior considering data imbalance,and the main work includes:(1)aiming at the problem that the sensitivity and reliability of the feature index items are difficult to achieve the expected abnormal identification effect due to the unbalanced data in the abnormal electricity consumption identification method and the large feature space corresponding to the user’s electricity consumption behavior characteristics,and propose an abnormal electricity consumption identification method based on feature optimization integrated learning.By constructing an ensemble learning model based on feature optimization and solving it,simulation experiments verify the effectiveness of the proposed method for abnormal power consumption identification.(2)Aiming at the problem that the original collected data lacks a large number of data labels and cannot collect enough historical electricity consumption data,resulting in insufficient model training data,an abnormal electricity use identification method based on feature optimization transfer learning is proposed,and the effectiveness of the proposed method for abnormal electricity use identification is verified by simulation experiments by constructing a TCA_SVM recognition model based on feature optimization transfer learning.The work carried out in this paper deepens the research on the identification of abnormal electricity consumption behavior considering data imbalance,which has reference significance for power grid management.