Using ANN to Forecast Transportation Sector's Energy Consumption in Taiwan Based on Oil and Gas Price

摘要

In this study, four types of artificial neural network (ANN) were adopted to forecast transportation sector's energy consumption (TSEC) taking different number of input variables. By taking premium gasoline price (PGP), premium diesel oil price (PDOP), fuel oil price (FOP), raw material natural gas price (RMNGP), and fuel natural gas price (FNGP) as input variables, ANN could successfully forecast TSEC, the best mean absolute percentage error, mean square error, root mean square error, and correlation coefficient for training and testing were 15.03 % versus 24.43 %, 2792036.59 versus 11982081.08, 1670.94 versus 3461.51, and 0.71 versus 0.51, respectively.