Ensemble-Learning-Based Decision Support System for Energy-Theft Detection in Smart-Grid Environment

Theft of electricity poses a significant risk to the public and is most costly non-technical loss for an electrical supplier. In addition affecting quality energy supply strain on power grid, fraudulent use drives up prices honest customers creates ripple effect economy. Using data-analysis tools, smart grids may drastically reduce this waste. Smart-grid technology produces much information, including consumers’ unique electricity-use patterns. By analyzing machine-learning deep-learning methods successfully pinpoint those who engage in theft. This study presents ensemble-learning-based system detecting theft using hybrid approach. The proposed approach uses machine-learning-based ensemble model based majority voting strategy. work aims develop smart-grid information-security decision support system. employed theft-detection dataset facilitate automatic recognition environment (TDD2022). consists six separate thefts. experiments are performed four different scenarios. obtained results all highest accuracy 88%, 87.24%, 94.75%, 94.70% with seven classes consumer type, excluding type. suggested outperforms existing techniques terms when methodology compared state-of-the-art approaches.