A Machine Learning Based Day-Ahead Electricity Demand Firecasting

The abstract will be submitted with the final report. Since the website did not allow me to submit the result file of 18th Jan., I am writing a brief temporary abstract to be allowed to submit the result file. Two Light Gradient Boosting machine learning models are used to forecast the demand.The majority of the training set is used to train the first Light Gradient Boosting model, while only a small portion of that is used to correct the first model.