Re: 'trend value' formula needed


Andy Smith wrote:

What you need to do is to fit a plane to your 3 data points (or a least square fit if you have more data points).

This assumes that there is an (approximate) linear relationship between
the variables, and I don't see any indication in the question that this
is known to be the case. If it's not the case then a best-fit plane
could give hopelessly incorrect predictions.


Your model is C = aA + bB + k

- with three sets of (A,B,C) you have 3 simultaneous equations to solve for a,b and k; then you just use the formula above to read off C for a given A,B value.

.

Quantcast