I have a problem and I've no idea about how to do it.
I have a DB with 30,000 rows. Each row have some fields with numbers. Each number is the result of a test (like medical test with blood, pressure, height, size, age....) and the the last field is the global result: 0 Not OK, 1 OK.
I want to know which is the bets combination of the fields numbers to get the "OK". Actually in the DB, is about 90,8% ->1 and 9,2%-> 0. I want to know how can I get 99%->1 and 1%->0.
I tried to use Regression, but the % of my test are similar (9-10% not OK) and I don't know how to use Neuronal Networks because the output I don't know how to use :S
My idea is to get some formula as (this data is random):
Y = 0,59X +0,48Z -1,09A
And if Y> 0,2 I know that the result is going to be "1"