Developing Simple Models for Analysing Energy Performance of Buildings at Neighbourhood Scale Using Geographic Information Systems (GIS)

These results are part of a PhD thesis. The dataset includes the outcomes of 24 machine-learning models developed using a direct regression approach. The spreadsheets present the results for three algorithms: Random Forest, XGBoost, and Extra Trees, both with optimised parameters and default settings. The first spreadsheet corresponds to the Random Forest results, the second to XGBoost, and the third to Extra Trees.