A process I am running includes - as part of a series of formulas - one step that puts raw_score through a logistic function: raw_exp=EXP([raw_score])/(1+EXP([raw_score])). This process previously was run on raw_scores that ranged from -200 to -2, and worked fine.
However, the formula behind raw_score has now changed, and its output now ranges from 25 to as high as 12000. This poses an issue when reaching the raw_exp function, because apparently for any values of raw_score larger 709.7 or so EXP(raw_score) returns a null value instead of the actual result, which throws off the distribution of the final score downstream.
Is there any way I can get Alteryx to handle values larger than 709.7 in the EXP command normally? Alternatively, is there a tool I'm missing that would do the same thing the raw_exp formula is doing without the potential for errors, or a potential mathematical brute-force to make every raw_score value be lower than 709.7 (i.e. multiplying them all by 0.01)? The result of raw_exp gets rescaled to be from 1 to 100 anyway, so I don't think multiplying raw_score by 0.01 will have any mathematical downsides since I'd be applying it to the entire universe of scores, but I may be missing something.
