Machine Learning Models and Intra-Daily Market Information for the Prediction of Italian Electricity Prices

In this paper we assess how intra-day electricity prices can improve the prediction of zonal day-ahead wholesale in Italy. We consider linear autoregressive models with exogenous variables (ARX) and without interactions among predictors, non-parametric taken from machine learning literature. particular, implement Random Forests support vector machines, which should automatically capture relevant predictors. Given large number ARX are also estimated using LASSO regularization, improves predictions when regressors many selects important variables. addition to prices, predictors include official demand forecasts wind generation expectations. Our results show that performance simple model is mostly superior those models. The analysis relevance variables, variable importance measures, reveals market information successfully contributes forecasting performance, although impact differs