Abstract: This paper presents a novel approach for predicting residential heating demand by combining time-series clustering with deep learning models. Using a dataset of 4,660 district heating ...