Power Demand Prediction of North Cyprus using Machine Learning Regressor Models

The increase in population and the usage of a variety of electrical appliances and devices cause a continuous increase in the country’s power demand. By predicting future power demand, countries could know at which point they could face a lack of production and take precautions before having such a case. Machine learning regressors can help us get future predictions by fitting historical data using them. In this study, we obtained 19 years of historical power demand data from North Cyprus Electricity Authority’s official website and obtained different datasets which consist of different time intervals of 12, 24, 36, 48, 60, and 120 months as input and the following month power demand as output. Five different machine learning regressor models were compared according to their coefficient of determination (R
²
). The highest R
²
value was obtained by Linear Regression with 120-month input data.