This thesis investigates the predictive performance of Recurrent Neural Networks (RNNs) in forecasting excess return in zero-coupon bonds. We evaluate their performance using data from the U.S. and German bond markets. The ...