I wonder if anyone here knows of a good site for me to take a look at re: computational methods for least squares approximation. I'm not attempting anything too fancy, i.e. just a simple linear and perhaps a quadratic approach. Basically, I want to take the output from 5 precision resistor measurements and determine the best compensation curve for errors induced by other components in the circuit.

Another option is to simply do point-to-point linear interpolations based on the precision resistance values. As long as I choose those values carefully, perhaps even a linear point-to-point approach would do well enough to approximate the curve I currently see. But I would probably prefer a more comprehensive method like least squares. Any help is appreciated!