Casio fx-4200P

(Power
Regression)
The
calculator
uses
the
formula
y=A+x®
(iny=InA+
Binx)
for
power
regression.
*Data Input
Press
fwd
(Z
to
enter
the
LR
mode.
@
Press
GA
to
clear
the
statistical
memories.
@
Enter
each
data
item,
using
the
following
sequence:
Tn)
(X-DATA)
GH]
fin)
(Y-DATA)
(7).
*You
can
enter
identical
data
items
using
the
same
two
procedures
described
for
tinear
regression,
but
you
must
always
press
the
{in}
key
before
entering
both
X-DATA
and
Y-DATA.
*Deleting
Entered
Data
To
delete
entered
data,
use
the
same
procedures
described
for
linear
regression,
but
‘remember
to
always
press
the
{in}
key
before
entering
X-DATA
and
Y-DATA.
To
perform
power
regression
on the above data to deter-
«Calculations
mine
the
r
F
The
following
table
shows
the
results
produced
by
each
key
operation:
egression
formula
and
the
correlation
coeffi-
Key
Operation
Result
pale?
Then
use
the
regres-
=
sion
formula
to
estimate
»
feria)
a
Sn
EW)
Constant
term
A
W=-@
when
xi=
40
and
¢
when
Be
Regression
coefficient
B
(8)
=@)
yi=
1000
fea)
aca
Correlation
coefficient
r
o-@
In
Ge
FIED
EKIES|
Estimated
value
of
x
@)=63
fia)
x
GH
CEM
ETHDEY|
Estimated
value
of
y
@-8
{Constant term A)
(HF) (E>) Gn)
(A) Ge
{Regression coefficient B)
(Correlation coett
shiF T
(* when
francs
cient r)
Gy
aye 40)
In) 40 Gai} 7} Ge GED
{Ans} (Exe}
(¥ when yu
1000)
ill
Mma
Kas) (BE)
Example Operation Display
(uo (2) > “LR”
(in) 28 Ain) 241067 3.33220451
(in) 30 G9 (+) fin} 3033 by) 3.401197382
(in) 33 GFF} fin) 3895 OF 3.496507561
1 (19) 35 Ge) [in) 4491 7) 3.555348061
| (in} 38 HC Gin) 5717 Gy 3.63758616
0.2388010963
2.771866138
0.9989062545
6587.674882
20.26225639
If
we
assume
that
Iny=y,
InA=a’
and
Inx=.x,
the
power
regression
formula
poArx®
(iny=InA
+
Binx)
becomes
the
linear
regression
formula
y
=a+bx.
Therefore,
the
for-
mulas
for
constant
term
A,
regression
coefficient
8
and
correlation
coefficient
r
are
identi-
cal
for
power
and
linear
regression.
A
number
of
power
regression
calculation
results
differ
from
those
produced
by
linear
regression,
Note
the
following:
Linear
regression
|
Power
regression
<x
Linx
Lx?
L(Inx)?
Ly,
Liny
P
Ly
L(iny)
Exy
Linxlny
—53-