1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
|
#### Tests for covRob() -*- R -*-
#### ------------------
#### Original :
##
## Author: Kjell Konis
## Date : 9/1/2000
##
{
## Set seed here, as several of the algorithms do random subsampling:
set.seed(912)
data(stack.dat) == "stack.dat"
}
### NOTE: "mcd" is now "fastMCD" from robustbase, see ../../man/covRob.Rd
## -----> different values below
{
## A. test MCD estimator
temp <- covRob(stack.dat, estim = "mcd")
covmat <- c(60.9301308800237, 67.3032825008078, 25.1599419754581, 48.989225920203,
67.3032825008078, 77.3182350477543, 24.0218791860323, 54.9071524252168,
25.1599419754581, 24.0218791860323, 18.2790392640071, 18.4015998720991,
48.989225920203, 54.9071524252168, 18.4015998720991, 99.0289713383374)
locvec <- c(13.1538461538462, 56.1538461538462, 20.2307692307692, 85.3846153846154)
all(all.equal(as.vector(temp$cov), covmat),
all.equal(as.vector(temp$center), locvec))
}
{
## A1. test MCD estimator: correlation matrix
temp <- covRob(stack.dat, estim = "mcd", corr = TRUE)
cormat <- c(1, 0.980571405623223, 0.753905474267168, 0.63067176442357,
0.980571405623223, 1, 0.638983037772476, 0.627490017608493,
0.753905474267168, 0.638983037772476, 1, 0.432511606499819,
0.63067176442357, 0.627490017608493, 0.432511606499819, 1)
## locvec remains the same!
all(all.equal(as.vector(temp$cov), cormat),
all.equal(unname(temp$cov), cov2cor(matrix(covmat,4,4)), tol = 1e-14),
all.equal(as.vector(temp$center), locvec))
}
{
## B. test Donoho-Stahel estimator
temp <- covRob(stack.dat, estim = "donostah")
covmat <- c(28.1722973969717, 29.4143954121706, 12.02968525837, 17.6045582618356,
29.4143954121706, 33.9317506099465, 11.2031930724940, 22.1352308669148,
12.02968525837, 11.2031930724940, 8.29767048720646, 8.79402368139085,
17.6045582618356, 22.1352308669148, 8.79402368139085, 37.8869805892866)
locvec <- c(13.7290122612456, 56.9150778381848, 20.4336876111483, 86.286178380903)
all(all.equal(as.vector(temp$cov), covmat),
all.equal(as.vector(temp$center), locvec))
}
{
## B1. test Donoho-Stahel estimator: correlation matrix
temp <- covRob(stack.dat, estim = "donostah", corr = TRUE)
cormat <- c(1, 0.951361694176807, 0.786801248488209, 0.538851383472141,
0.951361694176807, 1, 0.66766805120483, 0.617356357773397,
0.786801248488209, 0.66766805120483, 1, 0.495980430114247,
0.538851383472141, 0.617356357773397, 0.495980430114247, 1)
## locvec remains the same!
all(all.equal(as.vector(temp$cov), cormat),
all.equal(unname(temp$cov), cov2cor(matrix(covmat,4,4)), tol = 1e-14),
all.equal(as.vector(temp$center), locvec))
}
{
## C. test M estimator
temp <- covRob(stack.dat, estim = "M")
covmat <- c(35.7213914866887, 39.4577669180549, 14.7504711397620, 28.7208199367112,
39.4577669180549, 45.3292140244877, 14.0832612413037, 32.1903114086941,
14.7504711397620, 14.0832612413037, 10.7164174460066, 10.7882708196867,
28.7208199367112, 32.1903114086941, 10.7882708196867, 58.0575259335376)
locvec <- c(13.1538461538462, 56.1538461538462, 20.2307692307692, 85.3846153846154)
all(all.equal(as.vector(temp$cov), covmat),
all.equal(as.vector(temp$center), locvec))
}
{
## C1. test M estimator: correaltion matrix
temp <- covRob(stack.dat, estim = "M", corr = TRUE)
cormat <- c(1, 0.980571405623223, 0.753905474267168, 0.63067176442357,
0.980571405623223, 1, 0.638983037772476, 0.627490017608493,
0.753905474267168, 0.638983037772476, 1, 0.432511606499819,
0.63067176442357, 0.627490017608493, 0.432511606499819, 1)
## locvec remains the same!
all(all.equal(as.vector(temp$cov), cormat),
all.equal(unname(temp$cov), cov2cor(matrix(covmat,4,4)), tol = 1e-14),
all.equal(as.vector(temp$center), locvec))
}
{
## D. test quadrant-correlation estimator
temp <- covRob(stack.dat, estim = "pairwiseqc")
covmat <- c(30.6305178882016, 22.2434382439422, 11.2069103829590, 9.82255239190555,
22.2434382439422, 22.3371486310289, 7.65869436281274, 10.9896726674399,
11.2069103829590, 7.65869436281274, 6.77696481158792, 4.64292372384087,
9.82255239190555, 10.9896726674399, 4.64292372384087, 22.0645365958676)
locvec <- c(14.3125,57.125,20.3125,86.375)
all(all.equal(as.vector(temp$cov), covmat),
all.equal(as.vector(temp$center), locvec))
}
{
## D1. test quadrant-correlation estimator: correlation matrix
temp <- covRob(stack.dat, estim = "pairwiseqc", corr = TRUE)
cormat <- c(1,0.850375778589203,0.777841506522661,0.377832906137698,
0.850375778589203,1,0.622476923770166,0.495020842047739,
0.777841506522661,0.622476923770166,1,0.379687699883006,
0.377832906137698,0.495020842047739,0.379687699883006,1)
## locvec remains the same!
all(all.equal(as.vector(temp$cov), cormat),
all.equal(unname(temp$cov), cov2cor(matrix(covmat,4,4)), tol = 1e-14),
all.equal(as.vector(temp$center), locvec))
}
{
## E. test GK estimator
temp <- covRob(stack.dat, estim = "pairwisegk")
covmat <- c(17.0869329181457, 17.4702935925913, 7.57137332030014, 9.24173054467015,
17.4702935925913, 20.7014764200611, 6.3802169389871, 10.8162476004288,
7.57137332030014, 6.3802169389871, 5.81430356242458, 4.49536028962968,
9.24173054467015, 10.8162476004288, 4.49536028962968, 22.1345151316791)
locvec <- c(13.4,56.8,20.0666666666667,86.3333333333333)
all(all.equal(as.vector(temp$cov), covmat),
all.equal(as.vector(temp$center), locvec))
}
{
## E1. test GK estimator: correlation matrix
temp <- covRob(stack.dat, estim = "pairwisegk", corr = TRUE)
cormat <- c(1,0.928896841428589,0.759615384615385,0.475210729999254,
0.928896841428589,1,0.581548492752874,0.50529029133178,
0.759615384615385,0.581548492752874,1,0.396260375914721,
0.475210729999254,0.50529029133178,0.396260375914721,1)
## locvec remains the same!
all(all.equal(as.vector(temp$cov), cormat),
all.equal(unname(temp$cov), cov2cor(matrix(covmat,4,4)), tol = 1e-14),
all.equal(as.vector(temp$center), locvec))
}
{
## F. clean up
rm(temp, covmat, cormat, locvec)
TRUE
}
|
