Dear friends, i'm working on finding a way to hit the lottery jackpot.
The lottery i'm playing is a 90/6 (6 numbers of 90), with a total of 622.614.630 possible combinations.
What i have done so far, is:
1) Convert the lottery drawing to the combination number
2) Convert that number on a sequence of binary numbers (0-1-0-0-1-1-1 etc..), 30 numbers.
Now, in order to hit the jacpkpot, we have to forecast the correct binary sequence.
3) The binary sequence is splitted in 30 columns, and for each column we have a sequence from the past lottery drawings.
So what we have now are 30 binary sequences that are very long (a thousand numbers in each), and our plan is to forecast the next binary number for each of them (either a "1" or a "0").
Playing randomly, we have 50% chance of hit the correct number for each 30 sequences, but what we want is forecast correctly the next binary number of each of all 30 sequences.
I am using some neural networks:
input: 19 past binary numbers.
hidden neurons: 4
output: 1 binary number
Training rows: 100.
So far results seems decent, today hitten 20 correct numbers out of 30, but still not enough!
I'm using easy-nn plus (demo version) that only have max 100 training rows capabilities, and i was wondering if there is some free software out there that is easy to use and do not have any limitations?
Any input/help/idea is appreciated!
Thanks
P.s.
If you think i'm crazy, you are right