Saturday, October 13, 2007

Measuring Forecasting Error


Forecasting error can tell us how accurately we predicted the future values. This is the formula we can use in calculating the error present in our models.
MAD stands for "Mean Absolute Deviation". Comparing the actual values to the forecasts can allow us to see how well we are doing in executing sales or production strategies, among other things. However, having just the deviation value can be confusing, especially when dealing with large numbers over various product & service classes. How can we make this error value more useful? We do this by finding the error in terms of percentages. Here is the formula we'll want to use:
MAPE stands for "Mean Absolute Percent Error". Now, lets take a quick look at at error calculations for a made up set of data. We are looking for the error in forecasting for MP3 Players sold.

This is just a quick example of how you could calculate error. I'm not going to do this now, but with Excel or other graphing software you could even make a visual representation of your error to track it over time. This may ultimately help an analyst or project determine the usefulness of their forecasting models.

The next part of the series will be trend adjustments in your forecasts. I look forward to seeing you then. Thanks for reading.
----Sincerely, Trevor Stasik.


To return to initial post about forecasting, click HERE.
To view the next blog entry in this series, click HERE.



View Trevor Stasik's profile on LinkedIn

Post a Comment