For educational purposes, it would be nice to have a Neutral Network Viewer connecting to the Model output of Neural Network, that graphically depicts the network, including the input and output neurons. For instance if 'Neurons in hidden layers' is user-defined as 4,4 and, based on the selected features and targets, the number of input and output neurons is 3 and 2, respectively, it would show something like this:
There could be check boxes for labels: the input neurons could be labelled with the names of the features, and the output(s) with the name(s) of the target and/or the classes/categories. If the model has data on its input, also the connectors could be labelled with the weights.
Expected behavior
For educational purposes, it would be nice to have a Neutral Network Viewer connecting to the Model output of Neural Network, that graphically depicts the network, including the input and output neurons. For instance if 'Neurons in hidden layers' is user-defined as 4,4 and, based on the selected features and targets, the number of input and output neurons is 3 and 2, respectively, it would show something like this:
There could be check boxes for labels: the input neurons could be labelled with the names of the features, and the output(s) with the name(s) of the target and/or the classes/categories. If the model has data on its input, also the connectors could be labelled with the weights.