Abstract
In today’s rapidly evolving digital landscape, Power Marketing Management Platforms (PMMPs) have emerged as essential tools for businesses seeking to optimize their marketing strategies. These platforms integrate advanced technologies and data analytics to deliver targeted campaigns, automate processes, and generate valuable insights into customer behavior. However, traditional marketing approaches often face challenges such as data silos, poor cross-channel integration, and a limited ability to adapt to shifting consumer behaviors. This research investigates the application of machine learning (ML) algorithms to enhance electricity consumption analysis and user engagement within PMMPs. A comprehensive framework is proposed, incorporating advanced analytical models for user segmentation, behavioral analysis, and consumption forecasting. The K-Means clustering algorithm segments users based on their electricity consumption patterns, enabling a more tailored service approach. To analyze temporal shifts in user behavior and forecast consumption trends, this research introduces the Dynamic Random Synthesized Light Gradient Boosting Machine (DR-LightGBM). This model delivers actionable insights that support grid scheduling and resource optimization. The system develops personalized marketing strategies, such as dynamic pricing models and energy-saving incentives, to enhance user participation and promote energy conservation. The effectiveness of the proposed system is evaluated using several performance metrics. RMSE (38.11 kW), along with MAPE, MAE, and training time, assess the accuracy of consumption forecasting, while engagement rate and reduction in energy usage serve as business-relevant indicators of marketing impact. The experimental results highlight the significant potential of machine learning-driven approaches to improve user experience and operational efficiency in modern PMMPs.
Keywords
Get full access to this article
View all access options for this article.