Forecatsing using regression or neural networks

I need a small project to be completed. It is a load forecasting project and it can be done using either regression or artificial neural networks. The project involves first analyzing transportation data (distances, trip duration, day of week, time etc) . The complete data set would be provided. The data set would have comprehensive information on everything you need to complete the project. I would provide full data set along with the code book which has the variables in the data set. After analyzing the data, based on vehicle class (sedan, mid size, suv etc), different distances would have different energy requirements. 3 different energy consumption would be used and depending on distance and energy consumption, the total energy required for a day can be found. Using historical data, I need the model to predict distance, and energy requirements. The inputs to the neural network shall be: day, type of car, driving time. The influencing parameter would be driving distance and the output shall be energy consumption forecast. The attached paper provides clear information on data, parameters etc. The paper uses probability to forecast, this project involved neural networks. The forecast need not be as detailed as mentioned in the paper. Fig 1 and Fig 3 are key figures that would be needed by the neural network to predict.