DM41X Curve Fitting

The Advantage Pac CFIT program requires all data values to be positive/non-zero

This is quite limiting and perhaps a better method is the PPC ROM's CV program.

CV does the same curve fitting (including selecting the best fit ) without this restriction for curves where negative co-ordinates are appropriate.

It also gives the function of calculating an estimated x value from an entered y, which CFIT doesn't do.

Advantage Pac functions for statistical curve fitting.

CFIT program provides the functions to fit statistics data to LIN, Exp, LOG, Power and Best Fit

XEQALPHACFITALPHA

Clear Statistics Registers with C (CLΣ)
Enter Statistics in usual way with Y, ENTER, X, A(Σ+)
Don't use any data points ≤0
When all data pairs are entered press E (FIT)
Now choose which type of fit to use
- L = Linear
- LOG = Logarithmic
- Ex = Exponential
- P = Power
- B = Best Fit
R/S to cycle through the coefficients of the curve fit a, b, R^2 (error ratio)
You can next enter an x value to find the curve-fitted y value
pressing R/S can exit back to the Curves menu L LOG Ex P B to try a different fit to the same data
pressing R/S can exit back to the main menu to clear and enter a new set of data

X	0.72	1.31	1.95	2.58	3.14
Y	2.16	1.61	1.16	0.85	0.5

the curve found is y = 1.8515 - 1.1021 ln(x)

— John Pumford-Green 31/07/24 14:46 BST

Further Information