Electric load demand forecasting on Greek Energy Market using lightweight neural networks

In this paper, we deal with the Electric Load Demand Forecasting (ELDF) problem, considering the real case scenario of Greek Energy Market. ELDF constitutes a critical task accompanied by many applications, e.g., power systems operations and planning. In order to address the specific objectives and requirements of the Greek Energy Market, we propose a lightweight model with a novel loss function. We evaluate the effectiveness of the proposed model in terms of mean absolute percentage error, while we also evaluate its efficiency in terms of training/inference time, complexity, required memory, etc. The proposed model accomplishes the specific objectives dictated by the Greek Public Power Corporation, while it also achieves superior performance as compared to the successful baseline model ResNetPlus.