Advertisement
Table
4-5 describes
the models in more detail. When curve
fitting
uses
the linear model, the calculation is called linear
regressinn.
Calcula-
tions
for
the exPonential, logarithmic, and power i1odels
use
transformations
that allow the
data
to
be
ntt.a
by
standard linear
regression.
Table 4-5. Gurve
Fitting ilodels
Model
Equation
Transformation
Gonstraints
Linear
Logarithmic
Exponential
Power
y:B+Mx
y:
B
+
M lnx
Y:
BEM'
Y
--
BxM
None
y
:
B
+
Mlnx
y
versus
ln
x.
lny:
lnB
+
Mx
ln
y
versus
x.
lny
:lnB+Mlnx
ln
y
versus
ln
x
None
x
v
x>
o,y>
o
To
do curve fitting and
estimation:
1'
Enter
the data into two STAI
lists-one
for the r-values and
one
for the
y-values.-
The two items
in
an
x,A
pair must
have the
same
item number
in
each
list. Give both lists
meaningful
names. (You
do not need to name the current list, but
niming it
avoids possible confusion
later.)
2.
3.
4.
5.
6.
7.
From
the STAI
menu, press
ffi,
then
ffi.
PressffitodisplayamenuoflistnameS.T1oselectthex.
variable, press
the
aPpropriate menu kuy.
(If
the
current
list
is
unnamed, pressing
ffi
selects
it.)
Select
the
y-vanable
by
pressing
another menu key. The
;g19-1TS
displays the FRCST menu and the current model.
If
you need to_change models,
press
ffi
ffi,
and then
press
a menu
k.y
to
select
a
model.
To
-display
the curve fitting
results, press
ffi,
_
,
and/orffi.
To
do
estimation calculations:
1.
Kuy
in
the known value and
press
its menu
k"y.
2-
Press
the menu
k.y
for
the unknown.
4=
Statistics
Advertisement
