-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathInitializationCombinedGabor.R
202 lines (170 loc) · 8.42 KB
/
InitializationCombinedGabor.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
201
## Loading add-on packages
pack.names <- c("rpart","rpart.plot", "pROC", "caret", "RWeka", "ROCR")
require(plyr)
sapply(pack.names,library,character.only=TRUE)
sapply(pack.names,require,character.only=TRUE)
##Formula for decision tree with all image features
formula = as.formula("label ~ markov1 + markov2 + markov3 + markov4 + markov5 +
SDIntensityBG + IntensityDifference + avg.gabor.mean + avg.gabor.SD + Energy + Homogeneity + Entropy +
thirdordermoment + Inversevariance + Sumaverage + Variance + Clustertendency + MaxProbability +
Circularity + Compactness + Eccentricity + Solidity + Extent + RadialDistanceSD + SecondMoment +
Area + ConvexArea + Perimeter + ConvexPerimeter + EquivDiameter + MajorAxisLength +
MinorAxisLength")
label.selector <- function(x,index)
{
x <- as.vector(t(x))
value <- x[seq(0,4*(length(index)-1),4)+index]
return(value)
}
mode <- function(x)
{
tabSmpl<-tabulate(x)
ifelse((sum(tabSmpl == max(tabSmpl))>1),ceiling(mean(x)),which(tabSmpl== max(tabSmpl)))
}
rescale <- function(x)
{
value <- ifelse(x==1|x==2,1, ifelse( x==3, 2, 3))
return(value)
}
#1 and 2 are benign, 3 is malignant for ROC
rescale1 <- function(x)
{
value <- ifelse(x==1|x==2,0, 1)
return(value)
}
#1 is benign, 2&3 are malignant
rescale3 <- function(x)
{
value <- ifelse(x==2|x==3,1, 0)
return(value)
}
#Only works for this specific dataset
Avg.Gabor <- function( dataset )
{
gabor.features <- dataset[, 19:42]
avg.gabor.mean <- (gabor.features$gabormean_0_0 + gabor.features$gabormean_0_1 + gabor.features$gabormean_0_2 +
gabor.features$gabormean_1_0 + gabor.features$gabormean_1_1 + gabor.features$gabormean_1_2 +
gabor.features$gabormean_2_0 + gabor.features$gabormean_2_1 + gabor.features$gabormean_2_2 +
gabor.features$gabormean_3_0 + gabor.features$gabormean_3_1 + gabor.features$gabormean_3_2)/12
avg.gabor.SD <- (gabor.features$gaborSD_0_0 + gabor.features$gaborSD_0_1 + gabor.features$gaborSD_0_2
+ gabor.features$gaborSD_1_0 + gabor.features$gaborSD_1_1 + gabor.features$gaborSD_1_2
+ gabor.features$gaborSD_2_0 + gabor.features$gaborSD_2_1 + gabor.features$gaborSD_2_2
+ gabor.features$gaborSD_3_0 + gabor.features$gaborSD_3_1 + gabor.features$gaborSD_3_2)/12
avg.gabor.features <- data.frame(avg.gabor.mean, avg.gabor.SD)
return(avg.gabor.features)
}
##Only works for this specific dataset
bal_strat <- function(labels){
##Balance
ones <- which(labels[,4]==1) #201 24.8%
twos <- which(labels[,4]==2) #341 42.1%
threes <- which(labels[,4]==3) #268 33.1%
#twos will be slightly undersampled
#so that they don't represent more than
#40% of the cases or 324 total
#793 is new total case number
##Stratify 60% training, 30% testing, 10% validation
train.ones <- sample(201, 121, replace=FALSE)
train.twos <- sample(341, 205, replace=FALSE)
train.threes <- sample(268, 161, replace=FALSE)
train.index <- c(ones[train.ones], twos[train.twos], threes[train.threes])
test.ones <- sample(seq(1:201)[-train.ones], 60, replace=FALSE)
test.twos <- sample(seq(1:341)[-train.twos], 102, replace=FALSE)
test.threes <- sample(seq(1:268)[-train.threes], 80, replace=FALSE)
test.index <- c(ones[test.ones], twos[test.twos], threes[test.threes])
valid.ones <- sample(seq(1:201)[-c(train.ones, test.ones)], 20,replace=FALSE)
valid.twos <- sample(seq(1:341)[-c(train.twos, test.twos)], 34, replace=FALSE)
valid.threes <- sample(seq(1:268)[-c(train.threes, test.threes)], 27, replace=FALSE)
valid.index <- c(ones[valid.ones], twos[valid.twos], threes[valid.threes])
index = NULL
index$train = as.numeric(train.index)
index$test = as.numeric(test.index)
index$valid = as.numeric(valid.index)
return (index)
}
#Accuracies
calcacc <- function (results, index, g){
d = list(c("Train", "Test", "Valid"),
rep(c(paste("I", 1:4,sep = ""), paste("M", 1:4,sep = ""),
paste("A", 1:4,sep = "")), each =1))
table <- data.frame(data.frame(matrix(vector(), 3, 12,dimnames=d)))
ii=1
for(ii in 1:g){
if(ii < 5){
miss.iter <- which(results[,paste("I", ii, ".Pred", sep = "")]!=
results[,paste("I", ii, ".Label", sep = "")])
miss.mode <- which(results[,paste("I", ii, ".Pred", sep = "")]!=
results[,"Max.Mode"])
table["Train", paste("I", ii,sep = "")] <-
1-length(which(results[miss.iter, "Set"] == "train"))/length(index$train)
table["Test", paste("I", ii,sep = "")] <-
1-length(which(results[miss.iter, "Set"] == "test"))/length(index$test)
table["Valid", paste("I", ii,sep = "")] <-
1-length(which(results[miss.iter, "Set"] == "valid"))/length(index$valid)
table["Train", paste("M", ii,sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "train"))/length(index$train)
table["Test", paste("M", ii,sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "test"))/length(index$test)
table["Valid", paste("M", ii,sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "valid"))/length(index$valid)
} else {
miss.mode <- which(results[,paste("A", (ii-4), ".Pred", sep = "")]!=
labels[,ii-4])
table["Train", paste("A", (ii-4),sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "train"))/length(index$train)
table["Test",paste("A", (ii-4),sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "test"))/length(index$test)
table["Valid", paste("A", (ii-4),sep = "")] <-
1-length(which(results[miss.mode, "Set"] == "valid"))/length(index$valid)
}
}
return (table)
}
ms = seq(10, 250, 20)
mb = seq(2, 60, 4)
tunecontrols = expand.grid("minsplit" = ms, "minbucket" = mb, "cp" = 0.01,
"maxcompete" = 4, "maxsurrogate" = 5,
"usesurrogate" = 2, "surrogatestyle" =0,
"maxdepth" =50)
#Controls
ics = list(rpart.control(minsplit = 170, minbucket= 6),
rpart.control(minsplit = 170, minbucket= 6),
rpart.control(minsplit = 50, minbucket= 6),
rpart.control(minsplit = 110, minbucket= 6),
rpart.control(minsplit = 250, minbucket= 58),
rpart.control(minsplit = 210, minbucket= 6),
rpart.control(minsplit = 130, minbucket= 6),
rpart.control(minsplit = 150, minbucket= 6))
ehcontrols = c(rpart.control(minsplit = 510, minbucket= 2, cp = 0.01),
rpart.control(minsplit = 510, minbucket= 2, cp = 0.01),
rpart.control(minsplit = 510, minbucket= 2, cp = 0.01),
rpart.control(minsplit = 510, minbucket= 2, cp = 0.01),
rpart.control(minsplit = 510, minbucket= 2, cp = 0.01))
avgacc <- function(allaccs) {
yiah <- data.frame(
data.frame(matrix(vector(), 6, 8,
dimnames=list(c("TrainAcc","TrainErr", "TestAcc","TestErr",
"ValidAcc", "ValidErr"), c ("I1","I2","I3","I4",
"A1", "A2", "A3", "A4")))))
for (i in 1:3){
yiah[(i*2)-1,"I1"] = summary(sapply(allaccs, function (x) x[i, "I1"]))["Mean"]
yiah[(i*2)-1,"I2"] = summary(sapply(allaccs, function (x) x[i, "I2"]))["Mean"]
yiah[(i*2)-1,"I3"] = summary(sapply(allaccs, function (x) x[i, "I3"]))["Mean"]
yiah[(i*2)-1,"I4"] = summary(sapply(allaccs, function (x) x[i, "I4"]))["Mean"]
yiah[(i*2)-1,"A1"] = summary(sapply(allaccs, function (x) x[i, "A1"]))["Mean"]
yiah[(i*2)-1,"A2"] = summary(sapply(allaccs, function (x) x[i, "A2"]))["Mean"]
yiah[(i*2)-1,"A3"] = summary(sapply(allaccs, function (x) x[i, "A3"]))["Mean"]
yiah[(i*2)-1,"A4"] = summary(sapply(allaccs, function (x) x[i, "A4"]))["Mean"]
yiah[i*2,"I1"] = aerror(sapply(allaccs, function (x) x[i, "I1"]))
yiah[i*2,"I2"] = aerror(sapply(allaccs, function (x) x[i, "I2"]))
yiah[i*2,"I3"] = aerror(sapply(allaccs, function (x) x[i, "I3"]))
yiah[i*2,"I4"] = aerror(sapply(allaccs, function (x) x[i, "I4"]))
yiah[i*2,"A1"] = aerror(sapply(allaccs, function (x) x[i, "A1"]))
yiah[i*2,"A2"] = aerror(sapply(allaccs, function (x) x[i, "A2"]))
yiah[i*2,"A3"] = aerror(sapply(allaccs, function (x) x[i, "A3"]))
yiah[i*2,"A4"] = aerror(sapply(allaccs, function (x) x[i, "A4"]))
}
return (yiah)
}
aerror <- function(x) {
qt(0.975,df=length(x)-1)*sd(x)/sqrt(length(x))}