Trying to prove to my father that lottery numbers cannot be predicted and that there is no pattern to be found.
This approach predicts and calculates the best next fit for each pick individually, by looking at the last numbers and checking which number fits best next.
The learning process always looks at ATTRIBUTES amount of numbers as attributes, with the ATTRIBUTES+1th number being the class.
So for example the following numbers
3-6-30-31-32-43,
2-22-24-33-37-46,
6-7-12-17-18-34,
14-16-21-28-31-39
with ATTRIBUTES = 2 would give these training vectors for the first pick:
[3, 2, 6], [2, 6, 14]