Arrears Risk Prediction Of Large Power Customers Based On Multi-scale Feature Extraction

As the domestic electricity market continues to expand,electricity tariff recovery faces certain risks.High-voltage customers use large amounts of electricity each month,have large electricity tariff,and have different ways of paying electricity tariff.Therefore,analyzing and predicting the arrears risk of high-voltage customers has always been an important subject in the power industry.Taking high-voltage customers from a certain area as an example,this paper establishes a model to predict the risk of customer arrears,to further guide the Power Grid Corp to formulate arrears risk management policies.The research contents are as follows:(1)Construction of unified high-voltage customers data set.Integrate the power company’s basic data of the marketing system and business data of business halls,and other business-related data.Perform pre-processing such as quality assessment,cleaning,transformation,and integration on the original multi-source heterogeneous data to establish a multi-dimensional data set indexed by the customer number.(2)Analyzes the factors affecting the recovery of electricity tariff,extracts the risk features from multiple scales,and conduct feature screening based on the predictive capability to form a set of indicator system related to the power customer arrears risk.(3)Based on the logistic regression algorithm,establish a model to predict the risk of customer arrears,perform model training and model evaluation using customer data.The model’s evaluation results show that the prediction accuracy,precision and recall rate are still relatively accurate when customer information is not comprehensive enough.(4)Based on the output results of the logistic regression algorithm,risk ratings and credit score calculations are performed for customers.Visualize the risk of customer arrears to form a density distribution map of arrears risk.Based on the results of the statistical analysis of the risk distribution under different factors,suggestions are provided to further guide the Power Grid Corp to formulate arrears risk management policies.