Source code
Licensed under GNU GENERAL PUBLIC LICENSE Version 3
Training a neural network to separate points
Place points on the plane by clicking the right or left mousebutton. Left button will create a positive class point, right button a negative class point. Click a point again to remove it. Drag a point to move it. Positive class is shown by red circles, while negative is shown as blue circles. The rings around the circles represent the classification returned by the neural network.
To specify the network structure enter the numbers of neurons in the hiddenlayers in comma separated form. like "20,30" for two hidden layers with 20 and 30 neurons. After changing the network structure you need to press the init button for it to take effect.
Output examples:
Back to main page