Net Before Learning
(note: n2 and n5 are bias nodes)
(note: red is positive, green is negative)
(note: the bottom graph represents error over time)
Neural Nets
PART II - Xor Demo
| << Previous | Main | Next >> |
Xor Demo
With a delta of 0, Apex Forecaster is able to correctly associate the inputs and outputs of an XOR function. A simple 3-tier net is used for this application.
Net Before Learning (note: n2 and n5 are bias nodes)
| Net After Learning (note: red is positive, green is negative)
|
| Select Input / Output Pairs
| |
| Before Learning: (0 xor 1 -> 0.44)
| Before Learning: (1 xor 0 -> 0.43)
|
| Learned 1000 cycles (average error = 0.09) (note: the bottom graph represents error over time)
| |
| After Learning: (0 xor 1 -> 0.95)
| After Learning: (1 xor 0 -> 0.95)
|
<Under Construction>
| << Previous | Main | Next >> |
Next: Stock Prediction