1. Feed-Forward Networks

Figure 2 shows a feed-forward network with one hidden layer. This network attempts to fit the model

\begin{displaymath}
f_k(x)= F\left\{w_{0k}^{(2)} + \sum_{j=1}^{J} w_{jk}^{(2)}F ...
... w_{0j}^{(1)}
+ \sum_{i=1}^{I} w_{ij}^{(1)}x_i \right)\right\}
\end{displaymath}

for the output unit yk. Feed-forward means that information can only flow forward from the input units to the first hidden layer, from the first hidden layer to the second hidden layer, and so on. Information cannot flow between the units of one layer.

Figure 2: Feed-forward network with one hidden layer.

796



Method and Data Technologies   MD*TECH Method and Data Technologies
  http://www.mdtech.de  mdtech@mdtech.de