I have a CSV file with many kinds of values, some of them are numerical whole numbers such as 20, 30, 35, 40 etc. we will call this column A. While others are numerical doubles such as 9.2, 7.1, 6, 9.6 etc. this is column B.
When Encog analyses and normalises it identifies column B as "not classes" and column A as "classes". I do not want column A to be classes as it prevents me from then loading the normalisation data and normalising values such as 21 because it only knows 20 not 21.
One solution to this problem is to change some of the values in the column A from X to X.1 ie some of the 20's to 20.1. Although this stops column A from being seen as a class it is not a good or permanent solution.
I would like a way to say:
if (analyst.Script.Field[x].Integer == true) { analyst.Script.Field[x].Class = false; }
I know this is not possible to do after the normalisation but is there some way to preempt it?
I have a CSV file with many kinds of values, some of them are numerical whole numbers such as 20, 30, 35, 40 etc. we will call this column A. While others are numerical doubles such as 9.2, 7.1, 6, 9.6 etc. this is column B.
When Encog analyses and normalises it identifies column B as "not classes" and column A as "classes". I do not want column A to be classes as it prevents me from then loading the normalisation data and normalising values such as 21 because it only knows 20 not 21.
One solution to this problem is to change some of the values in the column A from X to X.1 ie some of the 20's to 20.1. Although this stops column A from being seen as a class it is not a good or permanent solution.
I would like a way to say:
if (analyst.Script.Field[x].Integer == true) { analyst.Script.Field[x].Class = false; }
I know this is not possible to do after the normalisation but is there some way to preempt it?