Cluster Aggregated GAN (CAG): A Cluster-Based Hybrid Model for Appliance Pattern Generation
Researchers have developed a novel framework called Cluster Aggregated GAN (CAG) to generate synthetic appliance data for non-intrusive load monitoring. This hybrid model addresses limitations in existing methods by differentiating between intermittent and continuous appliances. For intermittent devices, CAG uses a clustering module to group similar patterns and assign dedicated generators, while continuous appliances are handled by an LSTM-based generator. Experiments show CAG outperforms baseline methods in realism, diversity, and training stability. AI
IMPACT This new model could improve the accuracy and privacy of energy research by generating more realistic synthetic appliance data.