Analysis of a Log-Linear Model for Forecasting Electricity Demand Based on Economic Growth in Colombia

This study analyses the relationship between the Gross Domestic Product and Electricity Demand in Colombia. The development of these forecasting models supports long-term electricity sector planning. The study used annual data from 2006 to 2023 to construct two regression models—one linear correlation and the other log-linear. Actual data for 2024 were used to validate the predictive capacity of both models. Indicators such as R2, MAE, RMSD, and MAPE were used to evaluate the fitting period. The results show that the log-linear model achieved greater accuracy, with an MAPE of 1,30% in the fit and an error of 0,58% in the validation of the 2024 data. This approach demonstrates the usefulness of incorporating logarithmic transformations in energy models to obtain a more robust fit between economic and energy variables.