at time i find the predict function predicts values over 100% based on historical data.
is there anything i can configure to ensure the predicted value does not go over 100%, ie cutoff at 100%, or set max value as 100% ?
basically want to limit the predict value not to go beyond a certain number
Check this out :
It's got everything you can do with the predict command, I don't see any way to limit the upper bound to 100 but you could always use eval on the resulting field and make a condition saying if > 100 then make it 100.
Thanks David, but it does not tell much about limiting the prediction value.
in my case , im trying to get prediction for max cpu, and it seems with 3 months of data, its predicting over 100% of cpu, which is not true, so anything over 100% i would like to eval it to 100.
is there a way to do it via predict function, else would it be ok to do it with eval..
Hey again @jiaqya, all available options for the predict commands are in the reference sheet. I went through it again and there is no max boundary that can be set which means the only way to avoid these weird over 100% predictions is to use predict followed by eval setting the maximum value. Let me know if you're not able to build the eval I'll help you out with it 🙂
Thanks for helping David, i was having trouble evaluating the fields.
the field is maxCPU
after predict i get a field called prediction(maxCPU)
i was not able to eval this field due to the nature of function in it, due to brackets.
i was trying below, didnt work, see if you can help.