Vernier Spectro Pro - Manual (Page 48)

Spectro Pro Reference
50 Spectro Pro
automatic curve fit discussed above, where the parameters in the fit
equations are determined automatically using a least-squares technique,
the Manual Curve Fit allows you to adjust the parameters by hand.
A manual fit is often appropriate for instructional purposes. By adjusting
parameters manually, students will learn how each affects the fit. A
manual fit will also allow you to adjust a curve to fit a data series in the
way you want, possibly ignoring certain stray points within the series.
To perform a manual fit, click once on a graph to make it active.
Optionally, select a region of the graph using the mouse. This selection
region is used only for calculating the mean square error of the fitted
function with respect to the data. If you do not plan on using the mean
square error value, then it is not necessary to select a region. Then, select
Manual Curve Fit from the Analyze menu. A new dialog box will open.
First, you must select or enter an equation. The faster method is to select
an equation from the General Equation menu. In the screen shown
above, the proportional relationship has been selected. You can also
enter your own relationship in the Equation field. The equation must
contain between one and five adjustable parameters (single upper- or
lower-case letters) and the horizontal axis variable is entered as x.
The values of each parameter may be adjusted in the Parameters fields.
You can either type in a new value directly, or you may use the up and
down arrows to increase or decrease the values using the mouse. Click,
or click and hold, on the desired arrow.
The
∆
button near each parameter allows you to set the adjustment
increment applied when the arrows are used. Clicking on the
∆
button
selects the Auto Delta box. If the Auto Delta box is checked (default is
unchecked) then the increment will be made smaller when the parameter
reaches a magnitude similar to the increment.
Using the manual fit can be very helpful in understanding fitted
functions, but it can also be frustrating. A random or poorly chosen set
of parameter values may result in a function that does not cross the
graphed region at all. Blind adjustment of the parameters will not often