October 2017
Beginner to intermediate
270 pages
7h
English
In this phase of the operation of the network, the data will be input in the first layer and will flow from each unit to the corresponding units in the following layers. Then it will be summed and passed through in the hidden layers, and finally processed by the output layer. This process is totally unidirectional, so we are avoiding any recursive complications in the data flow.
The feedforward mechanism, has in the MLP its counterpart in the training part of the modeling process, which will be in charge of improving the model's performance. The normally chosen algorithm is called backpropagation.
Read now
Unlock full access