Which error should I calculate?

[drir]

Hi,

I want to compare (statistically) different models which predict the values y at several x-values. Therefore I want to calculate the 'total error' between the exact (measured) y-values and the calculated y-values using different models. My problem is that I'm not sure which method to use to calculate the 'total error' for each model. Should I use the sum of squared errors, the sum of the absolute errors or some other technique?

Thanks

 

[magoo2]

If you're just interested in the error at a particular value of x, then the error is the difference between the exact y and the model value for y. If however you're interested in the set of points for the actual and the model results, then the rms (root mean squared) approach seems appropriate.

 

[SomptingGuy]

Your "exact" is just a model anyway, using unreliable measurement equipment.

- Steve

 

[drir]

Yes, that is true! But I use this measured data as the input for my model in order to see if it can represent this measured data so I assume that if it can do the error is zero!

 

[IRstuff]

Least squares is the typical approach, assuming that your model behavior is roughly linear.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[rb1957]

r^2 is usually the figure of merit (about how well an approximation (trend line) fits the data)