Cluster Aggregated GAN (CAG): A Cluster-Based Hybrid Model for Appliance Pattern Generation
Researchers propose a new generative model called Cluster Aggregated GAN (CAG) for creating synthetic data of appliance patterns. The model addresses limitations in existing approaches by treating intermittent and continuous appliances separately, using clustering to allocate dedicated generators for each device's unique behavior. This approach improves training stability and output fidelity compared to previous methods.
Researchers propose a new generative model called Cluster Aggregated GAN (CAG) for creating synthetic data of appliance patterns. The model addresses limitations in existing approaches by treating intermittent and continuous appliances separately, using clustering to allocate dedicated generators for each device's unique behavior. This approach improves training stability and output fidelity compared to previous methods.
---
Why it matters: This matters because it enables more accurate and detailed simulation of real-world energy consumption patterns, which is essential for developing non-intrusive load monitoring algorithms and ensuring privacy in energy research.
Source: https://arxiv.org/abs/2512.22287
This article was originally published at: https://arxiv.org/abs/2512.22287