WAL-M
LINEAR REGRESSION EQUATI
Annual Sales. Forecast
Fiscal $ million Error
Year X Data Forecast $ million
1986 0 $ 11,909 $ 1,925 $ 9,984
1987 1 $ 15,959 $ 11,624 $ 4,335
1988 2 $ 20,649 $ 21,323 $ (674)
1989 3 $ 25,811 $ 31,022 $ (5,211)
1990 4 $ 32,602 $ 40,721 $ (8,119)
1991 5 $ 43,887 $ 50,420 $ (6,533)
1992 6 $ 55,484 $ 60,120 $ (4,636)
1993 7 $ 67,344 $ 69,819 $ (2,475)
1994 8 $ 82,494 $ 79,518 $ 2,976
1995 9 $ 93,627 $ 89,217 $ 4,410
1996 10 $ 104,859 $ 98,916 $ 5,943
1997 11 $ 108,615
1998 12 $ 118,314
1999 13 $ 128,014
Average Error $ 0
Sum of Squares of Errors $ 345,934,011
Model's Standard Error of Estimate $ 6,199.8
Intercept. $ million 1924.77
Slope, $ million/year 9699.14
Coeff. of Correlation, R 0.98369
Coeff. of Determination, R^2 0.96765
LINEST Output for Linear Model
9699.14 1924.77
591.1246 3497.1403
0.96765 6199.8
269.22046 9
10,348,057,082 345,934,011
DATA VALUES, LINEAR REGRESSION LINE, AND MINIMUM AND MAXIMUM CO
$140,000
$120,000
ANNUAL SALES, $ MILLION
$100,000
$80,000
$60,000
$40,000
ANNUAL SALES, $ MILLI
$-
$20,000
$40,000
$60,000
$80,000
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
WAL-MART STORES, INC.
LINEAR REGRESSION EQUATION FITTED TO DATA FOR ANNUAL SALES
Std. Fcst 80% Confidence Limits
X-XM Error, SFE Minimum Maximum SCATTER PLOT WITH LI
years $ million $ million $ million
$120,000
ANNUAL SALES, $ MILLION
-5 $ 7,118 $ (7,920) $ 11,769
$100,000
-4 $ 6,894 $ 2,090 $ 21,158 f(x) = 9699.13636363636
-3 $ 6,714 $ 12,038 $ 30,609 $80,000 R² = 1
-2 $ 6,582 $ 21,918 $ 40,126 $60,000
-1 $ 6,502 $ 31,728 $ 49,714
$40,000
0 $ 6,475 $ 41,465 $ 59,376
1 $ 6,502 $ 51,127 $ 69,113 $20,000
2 $ 6,582 $ 60,715 $ 78,922 $-
3 $ 6,714 $ 70,232 $ 88,803 0 1 2 3 4
4 $ 6,894 $ 79,683 $ 98,751 X (X = 0 FOR 1986, 1 FOR
5 $ 7,118 $ 89,072 $ 108,761
6 $ 7,383 $ 98,404 $ 118,826
7 $ 7,685 $ 107,686 $ 128,942 ERR
8 $ 8,018 $ 116,924 $ 139,103 Note that the scatter
Therefore, the straigh
Student's t = 1.3830
$15,000
Model Specification:
FORECAST ERROR, $ MILLION
$10,000
Y = 1924.77 + 9699.14*X
contoh untuk tahun 1997, maka $5,000
Y1997 = 1924.77 + 9699.14*11
Y1997 = $ 108,615
$-
$(5,000)
$(10,000)
1986
1987
1988
1989
1990
D MINIMUM AND MAXIMUM CONFIDENCE LIMITS
1992
1993
1994
1995
1996
1997
1998
1999
CATTER PLOT WITH LINEAR TREND LINE
000
000
f(x) = 9699.13636363636 x + 1924.77272727271
000 R² = 1
000
000
000
$-
0 1 2 3 4 5 6 7 8 9 10
X (X = 0 FOR 1986, 1 FOR 1987, ..., 10 FOR 1996)
ERROR PATERN
Note that the scatter about zero is NOT random.
Therefore, the straight line is NOT a valid model.
000
000
000
$-
000)
000)
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
WAL-MART STORES, INC.
QUADRATIC REGRESSION EQUATION FITTED TO DATA FOR ANNUAL SA
Annual Sales. Forecast Std. Fcst
Fiscal $ million Error X-XM Error, SFE
Year X X^2 Data Forecast $ million years $ million
1986 0 0 $ 11,909 $ 10,934 $ 975 -5 $ 2,450
1987 1 1 $ 15,959 $ 15,228 $ 731 -4 $ 2,373
1988 2 4 $ 20,649 $ 20,722 $ (73) -3 $ 2,311
1989 3 9 $ 25,811 $ 27,419 $ (1,608) -2 $ 2,266
1990 4 16 $ 32,602 $ 35,316 $ (2,714) -1 $ 2,238
1991 5 25 $ 43,887 $ 44,414 $ (527) 0 $ 2,229
1992 6 36 $ 55,484 $ 54,714 $ 770 1 $ 2,238
1993 7 49 $ 67,344 $ 66,215 $ 1,129 2 $ 2,266
1994 8 64 $ 82,494 $ 78,917 $ 3,577 3 $ 2,311
1995 9 81 $ 93,627 $ 92,821 $ 806 4 $ 2,373
1996 10 100 $ 104,859 $ 107,925 $ (3,066) 5 $ 2,450
1997 11 121 $ 124,231 6 $ 2,541
1998 12 144 $ 141,738 7 $ 2,645
1999 13 169 $ 160,446 8 $ 2,760
Average Error $ 0
LINEST Output for Linear Model Model Specifi
600.605 3693.087 10933.846154 Y= 10,933.85 + 3,69
72.853 756.407 1625.7769604 contoh untuk tahun 199
0.9966 2133.9779972 #N/A Y= 10,933.85 + 3,693,0
1170.1672 8 #N/A Y=
10657560196 36430897 #N/A
DATA VALUES, QUADRATIC REGRESSION LINE, AND MINIMUM AND MAXIMUM CO
LIMITS
$160,000
$140,000
ANNUAL SALES, $ MILLION
$120,000
$100,000
$80,000
$60,000
$40,000
$20,000
$-
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
A FOR ANNUAL SALES SCATTER PLOT WITH QUADRATIC TREND LINE
80% Confidence Limits $120,000
ANNUAL SALES, $ MILLION
Minimum Maximum
$100,000
f(x) = 600.6048951049 x² + 3693.0874125874 x + 10933.846153846
$ million $ million R² = 1
$ 7,512 $ 14,356 $80,000
$ 11,913 $ 18,542 $60,000
$ 17,494 $ 23,950
$40,000
$ 24,254 $ 30,583
$ 32,190 $ 38,442 $20,000
$ 41,301 $ 47,528 $-
$ 51,588 $ 57,840 0 1 2 3 4 5 6 7 8 9 10
$ 63,050 $ 69,380 X (X = 0 FOR 1986, 1 FOR 1987,..., 10 FOR 1996)
$ 75,689 $ 82,145
$ 89,506 $ 96,135
$ 104,503 $ 111,347 ERROR PATERN
$ 120,681 $ 127,781 The scatter about zero is random
$ 138,043 $ 145,433 The quadratic model is therefore valid
FORECAST ERROR, $ MILLION
$ 156,591 $ 164,301
$4,000
Student's t = 1.39681530974386 $3,000
$2,000
$1,000
Model Specification: $-
Y= 10,933.85 + 3,693,09*X + 600,605*X^2 $(1,000)
toh untuk tahun 1997, maka $(2,000)
$(3,000)
10,933.85 + 3,693,09*(11) + 600,605*(121) $(4,000)
$ 124,231
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
AND MAXIMUM CONFIDENCE
1994
1995
1996
1997
1998
1999
TREND LINE
x + 10933.846153846
6 7 8 9 10
10 FOR 1996)
ERN
o is random
therefore valid
1993
1994
1995
1996
WAL-MART STORES, INC.
EXPONENTIAL REGRESSION EQUATION FITTED TO DATA FOR ANNUAL SAL
Annual Sales. Forecast Error
Fiscal $ million Ratio. X-XM
Year X Data Forecast Data/Fost Forecast years
1986 0 $ 11,909 $ 13,153 0.9054 -9.46% -5
1987 1 $ 15,959 $ 16,443 0.9706 -2.94% -4
1988 2 $ 20,649 $ 20,556 1.0045 0.45% -3
1989 3 $ 25,811 $ 25,698 1.0044 0.44% -2
1990 4 $ 32,602 $ 32,126 1.0148 1.48% -1
1991 5 $ 43,887 $ 40,163 1.0927 9.27% 0
1992 6 $ 55,484 $ 50,209 1.1051 10.51% 1
1993 7 $ 67,344 $ 62,769 1.0729 7.29% 2
1994 8 $ 82,494 $ 78,471 1.0513 5.13% 3
1995 9 $ 93,627 $ 98,100 0.9544 -4.56% 4
1996 10 $ 104,859 $ 122,639 0.8550 -14.50% 5
1997 11 $ 153,317 6
1998 12 $ 191,669 7
1999 13 $ 239,614 8
Geometric mean error 1.0000
LOGEST Output for Exponential Model Model Specifica
1.2501 13152.81 Y = 13,152.81*(1.25
0.00797843 0.0472010046 contoh untuk tahun 1997, maka
0.98864 0.0837 Y = 13,152.81*(1.2501^11)
783.054 9 Y=
5.4830 0.06302
DATA VALUES, EXPONENTIAL REGRESSION LINE, AND MINIMUM AND MAX
$200,000
CONFIDENCE LIMITS
$180,000
ANNUAL SALES, $ MILLION
$160,000
$140,000
$120,000
$100,000
$80,000
$60,000
$40,000
$20,000
$-
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
ES, INC.
SCATTER PLOT WITH EXPONENTIAL TREND L
TED TO DATA FOR ANNUAL SALES
Std. Fcst 80% Confidence Limits $140,000
ANNUAL SALES, $ MILLION
Error, SFE Minimum Maximum $120,000
f(x) = 13152.8073907409 exp( 0.223261068
percent $ million $ million
$100,000 R² = 1
9.61% $ 11,405 $ 14,900
$80,000
9.30% $ 14,327 $ 18,559
9.06% $ 17,980 $ 23,132 $60,000
8.88% $ 22,541 $ 28,856 $40,000
8.78% $ 28,227 $ 36,026 $20,000
8.74% $ 35,308 $ 45,017
$-
8.78% $ 44,115 $ 56,304 0 1 2 3 4 5 6 7 8
8.88% $ 55,056 $ 70,482
X (X = 0 FOR 1986, 1 FOR 1987,..., 10 FOR 19
9.06% $ 68,636 $ 88,305
9.30% $ 85,476 $ 110,723
9.61% $ 106,344 $ 138,934 ERROR PATTERN
9.97% $ 132,187 $ 174,447 The scatter about 0% is NOT random
10.37% $ 164,174 $ 219,163 Therefore, the exponential model is NOT v
10.82% $ 203,749 $ 275,479 15.00%
Student's t = 1.383028738 10.00%
ERROR, PERCENT
5.00%
Model Specification: 0.00%
Y = 13,152.81*(1.2501^X) -5.00%
ontoh untuk tahun 1997, maka -10.00%
= 13,152.81*(1.2501^11) -15.00%
$ 153,317
1986
1987
1988
1989
1990
1991
1992
1993
N LINE, AND MINIMUM AND MAXIMUM
E LIMITS
1993 1994 1995 1996 1997 1998 1999
PONENTIAL TREND LINE
7409 exp( 0.223261068422338 x )
4 5 6 7 8 9 10
OR 1987,..., 10 FOR 1996)
TTERN
% is NOT random
ial model is NOT valid
1990
1991
1992
1993
1994
1995
1996
