Recurrent neural networks for time-series prediction

Recurrent neural networks have been used for time-series prediction with good results. In this dissertation we compare recurrent neural networks with time-delayed feed forward networks, feed forward networks and linear regression models to see which architecture that can make the most accurate predictions. The data used in all experiments is real-world sales data containing two kinds of segments: campaign segments and non-campaign segments. The task is to make predictions of sales under campaigns and with this in mind we evaluate if we can get more accurate predictions if we only use the campaign segments when modeling the data. Throughout the entire project the KDD process have been used to give a structured work-process. The results showed that recurrent network is not better than the other evaluated algorithms, in fact, the time-delayed feed forward neural network showed to give the best predictions. The results also showed that we could get more accurate predictions when only using information from campaign segments.