|Posted: June 24, 2009, 4:51 pm - IP Logged|
The data has to be a measurement of something for things like neural nets AI's to work. Lottery data is irrational, but if a section were to mimick real data the result should be predictable as the next step in the sequence. Maybe if there were a way to adjust the data to fit or eliminate the more irrational portions results would improve. The trouble is I don't know of a lottery software really using neural nets to test it on. The only one was made to do stock prediction and only did Pick-3 as a side line, didn't work out for me.
Would a method have to include some sort of machine learning or be more like an Expert System?
Backtracking held to one particular mindset or more adaptable?
Which lottery software program didn't work out for you?