Ahmed, SMadanian, SMirza, FZain, S2022-11-282022-11-2820212021ACIS 2020 Proceedings. 12. https://aisel.aisnet.org/acis2020/12https://hdl.handle.net/10292/15669Accurate prediction of natural gas consumption is of great importance for supply-demand balances and investments. This paper aims to utilize and compare the performance of multiple powerful machine learning algorithms to accurately predict the consumption of natural gas in Bahçeşehir, Istanbul. The utilized algorithms include Linear Regression, Random Forest Regression, Multilayer perceptron with back propagation (MLP) and gradient boosting (XGBoost). The algorithms were evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and R Squared to ensure consistency. The final results indicated that XGBoost outperformed MLP by 0.02, Forest Regression and Linear Regression by 0.04 Mean Absolute Error. XGBoost is highly scalable, efficiently reduces compute time and makes optimal use of memory which makes it a more suitable model for prediction. Accurate predictions reduce loss to the economy and ensure a balance between supply and demand.© 2020 by authors. This is an open-access article licensed under a Creative Commons Attribution-NonCommercial 3.0 New Zealand, which permits non-commercial use, distribution, and reproduction in any medium, provided the original authors and ACIS are credited.Forecasting; Machine learning; XGBoost; MLP; Natural gas consumptionConference ContributionOpenAccess