timaschew
commented
13 years ago - load configration file (topology, synapses, training) -> modal dialog with normalisation option and on the fly visualization with java gnu plot
- load predefined networks (built-in)
- set training algorithm (backprop, rprop) with parameters (learnrate, max iteration, min error, etc.), set init synapse algo
- set topology (click into jung graph in specific overplay region (input, hidden, output)
- configure amount of hidden layer
- configure standard prefs for neurons (activation function)
- configure bias
- visualize net error
- visualize dataset (train or test) in Java gnu plot
- visaualize neuron model in neuron jung view