Forecasting day ahead electricity price using ARMA methods
Tutor / director / evaluatorRego, Erik
Document typeBachelor thesis
Rights accessOpen Access
The main objective in this research is to predict day-ahead prices using ARMA. Box and Jenkins developed a regression model to identify, assess and diagnose dynamic time series models in which the time variable plays a key role. In the area of EPF, several studies have already been conducted to estimate prices by ARMA: in some cases ARMA is combined with GARCH using a wavelet transform, in others a hybrid model with Radial Basis Function Neural Networks (RBFN) or with Support Vector Regression (SVR) is performed, and also a single ARIMA or the addition of a Transfer Function and Dynamic Regression have been used.