Such measure are really helpful for us (analysts) not least to keep track of our progress, but it’s not that interresting for business.
You couldn’t convince anyone with such a 4 letter acronym. The only indicator that count here is the amount of money such forecasting can save versus the current model. That need some post processing to derive $ value from the model. Sometimes decreasing the RMSE is not saving money, quite the opposite.
For the analyst, the RMSE/MAE (Mean average error) or the correlation between forecast and the testing set can give insights if we are a bit wrong most of the time, or very wrong on some critical spots.
I like your website! Very interesting and informative posts and therefore original! I’m following yours now!
http://www.guccibagsforsale.com/Best-Gucci-wallet-233024-G-hot-sale_p262gucci bags tags
Tell me what you're thinking...
Email Address (required)
Speak your mind
Download Alexa toolbar
Copyright © 2008 · Revolution Code Blue theme by Brian Gardner
Get a Blog · WordPress · Log in