Derived Factors \(\mathfrak{f}_{t,DNN}^{(n),h}\) for Generated Using the Set of Yields

The figure above shows a 3D representation of \(\mathfrak{f}_{t,DNN}^{(n),h}\) generated by the MLP architecture using the set of yields \(\boldsymbol{Z}_t^{\text{y}} = \left\lbrace y_t^{(1/12)}, y_t^{(2/12)}, \ldots, y_t^{(60/12)} \right\rbrace\) in terms of maturity for the period of analysis. is a feedfoward neural networks architecture with 2 hidden layers (\(L=2\)), with \(16\) and \(4\) nodes respectively, and an output layer for each group of maturity \(n \in \left\lbrace 2,3,4,5 \right\rbrace\). Period of analysis ranges from 1993:01 to 2017:12.