Research on the Identification of Potential Harmful Behaviors in the Electricity Market Based on Cloud Model and STACKING Heterogeneous Models

This research addresses the formidable challenge of identifying and mitigating covert detrimental trading practices within the electricity market. An innovative detection approach is proposed, integrating the cloud model with the STACKING ensemble model to provide real-time alerts for potential threats during transactions. This study begins by characterizing the harmful behaviors of various market participants following the inception of the spot market, thereby constructing a robust framework for behavior identification. The cloud model is then utilized to delineate the features of these harmful actions, generating a comprehensive set of attributes based on indexed assessments. Finally, leveraging the STACKING model’s superior classification and predictive capabilities, this study examines exemplary cases to detect the presence of detrimental activities among the entities under surveillance.