Installation
There is no Installation necessary to Keep it as simple as possible. Just unzip the file and copy the content to a Folder you want to have the application there. Then start the exe-file.

Creation of a Perceptron Network
Just drag the mouse over the perceptron, you want to add to the network and press the left button. Keep it pressed, move the mouse over the drawing board and release the button, when the perceptron is at the location you want to have it.

How to connect perceptrons
The direction of the Impulse should always be from top to bottom. So it is useful to have all receptors at the top and all actor at the bottom of the network. Between These two layers it is possible to habve so much additonal Perceptron-layers as you want.
To connect two perceptrons, just klick on the arrow at the bottom of a Perceptron Symbol on the drawing board. Just perform a simple klick and release the button again. Move the button to the Perceptron you want to connect with the one you just clicked on.
Move the mouse-cursor above the top arrow og this Perceptron and execute a left-klick again. The perceptrons are now connected, shown by a line between them.

Create a Training script
For example if a Network as shown below is created and shall be trained now, the user has to set values for two receptors (which must be stimulated) and an expected result must be defined for the one actor used in this network. So here the values 1,0,0 must be entered in the inputfield in the Training console at the bottom of the Screen, when the first recpetor ist set to 1, the second is set to 0 and the expected value is also 0.
To define a complete NOT-Training-Script the four lines to be entered are these:

_1,0,0
0,1,0
1,1,0
0,0,1_

just enter this lines, press add line after each sequence, load the Network and press the start-button in the training console at the bottom of the Screen.

When the lines are entered and the network ist created, press the START-Button in the console at the bottom of the Screen in order to start a training-cycle. When the Network has fullfilled all requested functions, the Training process will stop automatically.

Last edited Dec 14, 2015 at 6:51 PM by bluecobra, version 2