Natural Log of the Determinant of the Covariance Matrix
SubMuli
5
4.86504
Upper
5
4.02474
Wilhelm
5
-5.77102
Pooled
5
4.30089
The SAS System
The DISCRIM Procedure
Test of Homogeneity of Within Covariance Matrices
Chi-Square
DF
Pr > ChiSq
49.126781
30
0.0153
Since the Chi-Square value is significant at the 0.1 level, the within covariance matrices will be used in the discriminant function. Reference: Morrison, D.F. (1976) Multivariate Statistical Methods p252.
The SAS System
The DISCRIM Procedure
Generalized Squared Distance to Zones
From Zones
SubMuli
Upper
Wilhelm
SubMuli
4.86504
16.42265
162.37251
Upper
18.93261
4.02474
123.32096
Wilhelm
10.52910
45.17244
-5.77102
The SAS System
The DISCRIM Procedure
Classification Summary for Calibration Data: WORK.CRUDEOILSAMPLE
Resubstitution Summary using Quadratic Discriminant Function
Number of Observations and Percent Classified into Zones
Characteristic Roots and Vectors of: E Inverse * H, where H = Type III SSCP Matrix for Zones E = Error SSCP Matrix
Characteristic Root
Percent
Characteristic Vector V'EV=1
vanadium
iron
beryllium
SathydroCarb
AromhydroCarb
4.17841435
86.25
0.04520460
-0.00835607
0.33510578
-0.06343120
-0.02721843
0.66601376
13.75
-0.01708629
0.00561389
0.27026635
-0.12953980
0.05543973
0.00000000
0.00
-0.00739870
0.00239697
0.38022080
0.02723223
-0.00331487
0.00000000
0.00
0.04500856
0.01198739
-0.01143913
-0.00580153
-0.01157736
0.00000000
0.00
0.04773022
-0.00454769
-0.05266021
0.09044752
0.01348588
MANOVA Test Criteria and F Approximations for the Hypothesis of No Overall Zones Effect H = Type III SSCP Matrix for Zones E = Error SSCP Matrix
S=2 M=1 N=23.5
Statistic
Value
F Value
Num DF
Den DF
Pr > F
Wilks' Lambda
0.11591099
18.98
10
98
<.0001
Pillai's Trace
1.20665556
15.21
10
100
<.0001
Hotelling-Lawley Trace
4.84442811
23.43
10
70.804
<.0001
Roy's Greatest Root
4.17841435
41.78
5
50
<.0001
NOTE: F Statistic for Roy's Greatest Root is an upper bound.
NOTE: F Statistic for Wilks' Lambda is exact.
Characteristic Roots and Vectors of: E Inverse * H, where H = Contrast SSCP Matrix for Linear Zones E = Error SSCP Matrix
Characteristic Root
Percent
Characteristic Vector V'EV=1
vanadium
iron
beryllium
SathydroCarb
AromhydroCarb
1.42921965
100.00
-0.04634853
0.01005049
-0.10386533
-0.02767747
0.05517612
0.00000000
0.00
0.01744996
-0.00427799
-0.25831367
0.16392059
-0.00759932
0.00000000
0.00
0.00495834
0.00102439
0.49592448
-0.01389307
0.00971742
0.00000000
0.00
0.06326601
0.00303710
0.04961471
-0.02253983
0.01576534
0.00000000
0.00
0.01463633
0.01191122
-0.08303571
0.03772290
-0.02638505
MANOVA Test Criteria and Exact F Statistics for the Hypothesis of No Overall Linear Zones Effect H = Contrast SSCP Matrix for Linear Zones E = Error SSCP Matrix
S=1 M=1.5 N=23.5
Statistic
Value
F Value
Num DF
Den DF
Pr > F
Wilks' Lambda
0.41165483
14.01
5
49
<.0001
Pillai's Trace
0.58834517
14.01
5
49
<.0001
Hotelling-Lawley Trace
1.42921965
14.01
5
49
<.0001
Roy's Greatest Root
1.42921965
14.01
5
49
<.0001
Characteristic Roots and Vectors of: E Inverse * H, where H = Contrast SSCP Matrix for Quadratic Zones E = Error SSCP Matrix
Characteristic Root
Percent
Characteristic Vector V'EV=1
vanadium
iron
beryllium
SathydroCarb
AromhydroCarb
3.49137381
100.00
0.04105271
-0.00711427
0.38100087
-0.08725933
-0.01599668
0.00000000
0.00
-0.02204002
0.00347824
0.16233885
-0.09071201
0.06233302
0.00000000
0.00
-0.00899377
0.00286002
0.39134067
0.02007714
0.00000000
0.00000000
0.00
0.04270061
-0.00839115
-0.09026910
0.11625030
0.00000000
0.00000000
0.00
0.05117388
0.01140885
0.00000000
0.00000000
0.00000000
MANOVA Test Criteria and Exact F Statistics for the Hypothesis of No Overall Quadratic Zones Effect H = Contrast SSCP Matrix for Quadratic Zones E = Error SSCP Matrix
S=1 M=1.5 N=23.5
Statistic
Value
F Value
Num DF
Den DF
Pr > F
Wilks' Lambda
0.22264903
34.22
5
49
<.0001
Pillai's Trace
0.77735097
34.22
5
49
<.0001
Hotelling-Lawley Trace
3.49137381
34.22
5
49
<.0001
Roy's Greatest Root
3.49137381
34.22
5
49
<.0001
The SAS System
The GLM Procedure
Least Squares Means
Adjustment for Multiple Comparisons: Tukey-Kramer
Zones
vanadium LSMEAN
LSMEAN Number
SubMuli
4.44545455
1
Upper
7.22631579
2
Wilhelm
3.22857143
3
Least Squares Means for effect Zones Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: vanadium
i/j
1
2
3
1
0.0002
0.3809
2
0.0002
<.0001
3
0.3809
<.0001
Zones
iron LSMEAN
LSMEAN Number
SubMuli
33.0909091
1
Upper
22.2526316
2
Wilhelm
43.5714286
3
Least Squares Means for effect Zones Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: iron
i/j
1
2
3
1
0.0023
0.0480
2
0.0023
<.0001
3
0.0480
<.0001
Zones
beryllium LSMEAN
LSMEAN Number
SubMuli
0.17090909
1
Upper
0.43210526
2
Wilhelm
0.11714286
3
Least Squares Means for effect Zones Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: beryllium
i/j
1
2
3
1
0.0290
0.9219
2
0.0290
0.0283
3
0.9219
0.0283
Zones
SathydroCarb LSMEAN
LSMEAN Number
SubMuli
6.56090909
1
Upper
4.65815789
2
Wilhelm
6.79571429
3
Least Squares Means for effect Zones Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: SathydroCarb
i/j
1
2
3
1
<.0001
0.8865
2
<.0001
<.0001
3
0.8865
<.0001
Zones
AromhydroCarb LSMEAN
LSMEAN Number
SubMuli
5.4836364
1
Upper
5.7678947
2
Wilhelm
11.5400000
3
Least Squares Means for effect Zones Pr > |t| for H0: LSMean(i)=LSMean(j) Dependent Variable: AromhydroCarb