Time Series Analysis by XGBoost Model for Future Prediction of Power Consumption

This research study tries to investigate the power consumption in Romania taking the concern of time series analysis. This time series analysis has been done to forecast the probable consumption of the power based on past data record of 5.5 years. Extreme Gradient Boosting (XGBoost) has been incurred in this dataset to analyze the power consumption profile throughout the years, since power consumption strategy can play a vital role for minimizing the critical condition in near future. From the data analysis it is observed that power consumption is relatively high from Monday to Friday in every week whether the overall power consumption variation throughout the year is almost same. The maximum power consumption recorded is 9615 MW with minimum of 3889 MW throughout the years of dataset where standard deviation is 1043.65 MW. From the data analysis it is seen that minimum RMSE recorded is 119.05 for the training dataset and 144.98 for the validation dataset.