-
Notifications
You must be signed in to change notification settings - Fork 0
/
gridsearch.R
200 lines (191 loc) · 6.97 KB
/
gridsearch.R
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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
mu <- 1
sigma <- 2
N <- 1e7
DGP <- function(N, mean=mu, sd=sigma) {
x <- rnorm(N, mean, sd)
cat(N, "normals, mean =", mean(x), " std deviation =", sqrt(var(x)), "\n")
x
}
loglik <- function(par) {
mu <- par[1]
sigma <- par[2]
sum(dnorm(x, mu, sigma, log=TRUE))
}
search1 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100) {
## serial computations
## Create a sequence of mu-s:
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
## and a sequence of sigmas:
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
## Make grid out of two sequences:
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
## and list all grid points
points <- seq(length=nrow(grid))
##
res <- lapply(points, function(i) loglik(grid[i,]))
##
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}
search2 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100,
cores=detectCores()) {
## use mclapply (not on windows!)
library(parallel)
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
points <- seq(length=nrow(grid))
##
res <- mclapply(points, function(i) loglik(grid[i,]), mc.cores=cores)
##
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}
search3 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100) {
## use socketCluster
library(parallel)
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
points <- seq(length=nrow(grid))
## -------------------------------------------------
cl <- makePSOCKcluster(c("localhost", "localhost", "localhost", "localhost"),
outfile="cluster_log.txt")
on.exit(stopCluster(cl), add=TRUE)
print(cl)
##
res <- parLapply(cl, points, function(i) loglik(grid[i,]))
## --------------------------------------------------
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}
search4 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100) {
## use socketCluster
library(parallel)
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
points <- seq(length=nrow(grid))
## -------------------------------------------------
cl <- makePSOCKcluster(c("localhost", "localhost", "localhost", "localhost"),
outfile="cluster_log.txt")
on.exit(stopCluster(cl), add=TRUE)
print(cl)
clusterExport(cl, c("x", "loglik"))
clusterExport(cl, c("grid"), envir=parent.frame(1))
##
res <- clusterApply(cl, points, function(i) loglik(grid[i,]))
## --------------------------------------------------
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}
search5 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100) {
## use socketCluster
library(parallel)
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
points <- seq(length=nrow(grid))
## -------------------------------------------------
eth0 <- system("ifconfig eth0", intern=TRUE)
eth0 <- grep("inet addr:", eth0, value=TRUE)
ip <- sub(".*addr:([^ ]+).*", "\\1", eth0)
cl <- makePSOCKcluster(c(rep("localhost", 4), rep("puyol", 8)),
user="otoomet",
homogeneous=FALSE,
master=ip,
outfile="cluster_log.txt")
on.exit(stopCluster(cl), add=TRUE)
print(cl)
clusterExport(cl, c("x", "loglik"))
##
res <- clusterApplyLB(cl, points, function(i) loglik(grid[i,]))
## --------------------------------------------------
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}
search6 <- function(mulim=c(-5,5),
sigmalim=c(0.001,5),
nGrid=100) {
## use socketCluster
library(parallel)
library(snow)
mus <- seq(from=mulim[1], to=mulim[2], length=nGrid)
sigmas <- seq(from=sigmalim[1], to=sigmalim[2], length=nGrid)
grid <- as.matrix(expand.grid(mu=mus, sigma=sigmas))
points <- seq(length=nrow(grid))
## -------------------------------------------------
cl <- makeMPIcluster(4, outfile="cluster_log.txt")
on.exit(stopCluster(cl), add=TRUE)
print(length(cl))
clusterExport(cl, c("x", "loglik"))
##
res <- clusterApply(cl, points, function(i) loglik(grid[i,]))
## --------------------------------------------------
res <- matrix(unlist(res), nGrid, nGrid)
dimnames(res) <- list(mu=formatC(mus, format="f", digits=4, width=7),
sigma=formatC(sigmas, format="f", digits=4, width=7))
##
best <- which(res==max(res), arr.ind=TRUE)
i <- best[1]
j <- best[2]
bestMu <- mus[i]
bestSigma <- sigmas[j]
cat("Maximum", res[i,j], "at mu =", bestMu, "and sigma =", bestSigma, "\n")
res
}