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
160
161
162
163
164
165
166
167
168
169
|
#### Testing lmRob() -*- R -*-
####
## Original
## author: Jeffrey Wang
## date : 08/09/2000
##
{
## Generate some data for loop test ##
mode(gen.data <- function(coeff, n = 100, eps = 0.1, sig = 3,
snr = 1/20, seed = 837)
{
set.seed(seed)
x <- cbind(rnorm(n, 1), rnorm(n, 1)^3, exp(rnorm(n, 1)))
ru <- runif(n)
n1 <- sum(ru < eps)
u <- numeric(n)
u[ru < eps] <- rnorm(n1, sd = sig/snr)
u[ru > eps] <- rnorm(n - n1, sd = sig)
data.frame(y = x %*% matrix(coeff, ncol = 1) + u,
x1 = x[, 1], x2 = x[, 2], x3 = x[, 3], x4 = rnorm(n, 1))
}
) == "function"
}
{
class(simu.dat <- gen.data(1:3)) == "data.frame"
}
{
## test S-estimates with random resampling ##
m <- lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
control = lmRob.control(estim = "initial",
initial.alg = "random"))
all.equal(unname(coef(m)),
c(1.659806131, 2.06709376, 2.879434355, -0.2756236906))
}
## {
## ## test S-estimates with genetic algorithm ##
## all.equal(as.vector(coef(lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
## control = lmRob.control(estim = "initial",
## initial.alg = "genetic",seed = 100)))),
## c(0.9202865, 2.046525, 3.063134, -0.2163211),
## tolerance = 1.0e-6)
## }
{
## test MM-estimates with weight (B,B) ##
mBB <- lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
control = lmRob.control(weight = c("Bisquare", "Bisquare"),
efficiency = 0.7, initial.alg = "random", final.alg = "m"))
all.equal(unname(coef(mBB)),
c(1.121617602, 2.028109705, 2.920919887, -0.03255785645))
}
{
## test MM-estimates with weight (B,O) ##
mBO <- lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
control = lmRob.control(weight = c("Bisquare", "Optimal"),
efficiency = 0.95, initial.alg = "random", final.alg = "m"))
all.equal(unname(coef(mBO)),
c(1.021358214, 2.040216606, 2.915863868, 0.05542195195))
}
{
## test MM-estimates with weight (O,B) ##
mOB <- lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
control = lmRob.control(weight = c("Optimal", "Bisquare"),
efficiency = 0.9,initial.alg = "random", final.alg = "m"))
all.equal(unname(coef(mOB)),
c(1.062536432, 2.035703683, 2.918117149, 0.01624240296))
}
{
## test MM-estimates with weight (O,O) ##
mOO <- lmRob(y~x1+x2+x3+x4-1, data = simu.dat,
control = lmRob.control(weight = c("Optimal","Optimal"),
efficiency = 0.85, initial.alg = "random",final.alg = "m"))
all.equal(as.vector(coef(mOO)),
c(1.020023715, 2.040035389, 2.91604064, 0.05466841575))
}
{
## test Robust Wald test ##
all.equal(anova(mOO, test = "RWald")[,"P(>Wald)"][2:4],
c(0, 0, 0.842332812))
}
{
## test Robust F test ##
all.equal(anova(mOO,test = "RF")[,"Pr(F)"][2:4],
c(0, 0, 0.845138356))
}
## {
## ## test REWLS with oilcity data ##
## tmp <- lmRob(Oil~Market, data = oilcity, control =
## lmRob.control(efficiency = 0.77,
## initial.alg = "random",final = "adaptive"))
## all.equal(as.vector(tmp$coef),
## c(-0.07813668, 0.8574827),
## tolerance = 1.0e-6)
## }
{
## test REWLS with coleman data ##
data(coleman, package = "robustbase")
mCM <- lmRob(Y ~ . , data = coleman,
control = lmRob.control(efficiency = 0.77,
initial.alg = "random", final = "adaptive"))
all.equal(unname(coef(mCM)),
c(29.7577177, -1.69854147, 0.0851182371,
0.666168644, 1.18399532, -4.06675281), tol = 1e-6)
}
{
## test REWLS with stack.loss data ##
data(stack.dat)
tmp <- lmRob(Loss~.-1, data = stack.dat, control =
lmRob.control(weight = "Bisquare", initial.alg = "random",
efficiency = 0.77, final.alg = "adaptive"))
all.equal(as.vector(tmp$coef),
c(0.6127073, 0.9676439, -0.473352),
tolerance = 1.0e-6)
}
{
## test robust "mixed" linear models with wagner data
## In the future, use
## data(wagnerGrowth, package = "robustbase")
source(system.file("datasets", "wagner.q",
package = "robust")) # wagnerGrowth
## 21 levels + 3 levels + 4 continuous :
tmp <- lmRob(y ~ Region + Period + ., data = wagnerGrowth)
all.equal(unname(coef(tmp)),
c(-58.48739738,
4.24094749, 28.95751724, 25.57747551, 22.72475947, -0.9850417527,
10.7973689, 23.58086125, 14.47839294, 14.22681835, 8.319455272,
10.35773846, 15.3466895, 10.36446368, 2.029283378, -8.077244089,
6.805266348, 12.66957858, 5.855703339, 3.350434134, -6.422418986,
8.761413075, 16.27819707,
1.130854624, 0.3911569697, 3.726122795, 2.790172641),
tol = 1e-5)
## now with non-default control :
tmp2 <- lmRob(y ~ Region + Period + ., data = wagnerGrowth,
control = lmRob.control(weight = "Bisquare",
efficiency = 0.77, final.alg = "adaptive"))
## FIXME ?: This seems completely platform(?) dependent
}
{
## test fast procedure for lmRob ##
data(stack.dat)
tmp <- lmRob(Loss~., data = stack.dat, control = lmRob.control(
estim = "initial", initial.alg = "Fast"))
all.equal(c(as.vector(tmp$coef), tmp$scale), c(-35.64108, 0.8458725,
0.4452125, -0.08965558, 1.837017), tolerance = 1e-05)
}
{
## remove function ###
rm(gen.data, simu.dat, .Random.seed, tmp, tmp2, m, mBB)
TRUE
}
|
