-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path2.classification.R
314 lines (254 loc) · 9.26 KB
/
2.classification.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
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
######################################################
# Pattern Recognition Project (Week 2)
# Name(s) Student Number
# Anne Tjallingii: 911024836020
# Diego Montiel: 880505580110
######################################################
# install.packages("tree")
# install.packages("randomForest")
# install.packages("gbm")
# install.packages("e1071")
library(gbm)
library(tree)
library(ISLR)
library(MASS)
library(class)
library(e1071)
library(randomForest)
data = read.table("expr4T.dat", header = TRUE)
d = data
data = d
# Last column contains tissue label
nn = ncol(data)-1
m = sapply(data[,1:nn],mean)
s = sapply(data[,1:nn],sd)
sm = s/m
#print the graphs of the m and sm
par(mfrow = c(2,2));plot(m);plot(sm);plot(m[m < 15]);plot(sm[sm < 1]);
minsm= 2.5
minm= 20
# In the following, we make sure that tissue label is kept;
# by giving it an artificial sufficiently high value for m and sm
m = c(m,minm+1)
sm = c(sm,minsm+1)
dim(data)
length(which(sm > minsm & m > minm))
data = data[,which(sm > minsm & m > minm)]
dim(data)
#Check if these two genes are in the new dataset
data[,"ENSG00000271043.1_MTRNR2L2"]
data[,"ENSG00000229344.1_RP5.857K21.7"]
#For getting the names of the tissues
unique(data[dim(data)[2]])
tissue1 = "brain_amygdala"
tissue2 = "brain_hippocampus"
data = data.frame(data)
mydat = data[which(data$tissue == tissue1|data$tissue==tissue2),]
mydat = droplevels(mydat)
mydat
#######################
# Classification tree
#######################
par(mfrow =c(1,1))
set.seed(2)
n = round(dim(mydat)[1]*.5)
train1 = sample(1:nrow(mydat), n)
mydat.test = mydat[-train1,]
tissue.test = mydat$tissue[-train1]
#Applying tree
tree.dataset1 = tree(tissue ~., mydat, subset = train1)
tree.pred = predict(tree.dataset1, mydat.test, type = "class")
tree.table = table(tree.pred, tissue.test)
tree.table
((tree.table[1] + tree.table[4]) / sum(tree.table))
plot(tree.dataset1)
text(tree.dataset1, pretty = 0)
tree.cv2 = cv.tree(tree.dataset1)
plot(tree.cv2$size, tree.cv2$dev, type = "b")
#optimal is 3 so prune back
prune.genes = prune.tree(tree.dataset1, best = 3)
plot(prune.genes)
text(prune.genes, pretty=0)
prediction.prune = predict(prune.genes, mydat.test, type = "class")
tree.table.prune = table(prediction.prune, tissue.test)
tree.table.prune
((tree.table[1] + tree.table[4]) / sum(tree.table))
#######################
# Regression tree
#######################
#fix(data)
n = round(dim(data)[1]*.5)
train2 = sample(1:nrow(data),n)
test2 = data[-train2,]
tree.dataset2 = tree(ENSG00000271043.1_MTRNR2L2 ~.-tissue, data = data, subset = train2)
par(mfrow = c(1,1))
summary(tree.dataset2)
plot(tree.dataset2); text(tree.dataset2, pretty = 0)
tree.cv = cv.tree(tree.dataset2)
plot(tree.cv$size, tree.cv$dev, type = "b")
#Prunning it doesnt make sense because the biggest value from the plot before
#is already taken in it = 7
yhat = predict(tree.dataset2, newdata = test2)
plot(yhat, test2$ENSG00000271043.1_MTRNR2L2); abline(0,1)
mean((yhat-test2$ENSG00000271043.1_MTRNR2L2)^2)
######################################
# RANDOM FOREST Regression
#####################################
install.packages("randomForest")
library(randomForest)
set.seed(1)
#The argument mtry=13 indicates that all 13 predictors
#should be considered for each split of the tree-in other words,
#that bagging should be done
train3 = sample(1:nrow(data), nrow(data)/2)
test3 = data[-train3,]
bag = randomForest(ENSG00000271043.1_MTRNR2L2 ~.-tissue, data = data, subset = train3,
importance = TRUE, ntree=1000, mtry = 19)
importance(bag)
yhat.bag = predict(bag, newdata = test3)
mean((yhat.bag - test3$ENSG00000271043.1_MTRNR2L2)^2)
plot(yhat.bag, test3$ENSG00000271043.1_MTRNR2L2); abline(0,1)
plot(bag)
######################################
# RANDOM FOREST CLASSIFICATION
#####################################
set.seed(2)
n2 = round(dim(mydat)[1]*.5)
train.random.forest2 = sample(1:nrow(mydat), n2)
test.random.forest2 = mydat[-train.random.forest2,]
tissue.test2 = mydat$tissue[-train.random.forest2]
tissue.test2
bag.classification = randomForest(tissue~., data = mydat, subset = train.random.forest2,
mtry = 19, importance = TRUE, ntree = 25)
yhat.bag2 = predict(bag.classification, test.random.forest2, type = "class")
yhat.bag2
tree.table2 = table(yhat.bag2, tissue.test2)
tree.table2
plot(yhat.bag2)
((tree.table2[1] + tree.table2[4]) / sum(tree.table2))
##############################################
# SVM RADIAL
##############################################
#install.packages("e1071")
library(e1071)
n = round(dim(data)[1]*.5)
svm.train = sample(1:nrow(data),n)
svm.test = data[-svm.train,]
svm.fit = svm(ENSG00000271043.1_MTRNR2L2 ~.-tissue, data = data[svm.train,], kernel ="radial")
summary(svm.fit)
svm.fit
plot(svm.fit, data[svm.train, ])
prediction.svm = predict(svm.fit, newdata = svm.test)
plot(prediction.svm, svm.test$ENSG00000271043.1_MTRNR2L2); abline(0,1)
mean((prediction.svm-svm.test$ENSG00000271043.1_MTRNR2L2)^2)
############################################
#SVM POLYNOMIAL
###########################################
svm.fit2 = svm(ENSG00000271043.1_MTRNR2L2 ~ . -tissue, data = mydat[svm.train,], kernel = "polynomial")
prediction.svm2 = predict(svm.fit2, newdata = svm.test)
plot(prediction.svm2, svm.test$ENSG00000271043.1_MTRNR2L2); abline(0,1)
mean((prediction.svm2-svm.test$ENSG00000271043.1_MTRNR2L2)^2)
#############################################
# Predicting using boosting (CLASSIFICATION)
#############################################
#install.packages("gbm")
library(gbm)
set.seed(1)
n = round(dim(mydat)[1]*.5)
train = sample(1:nrow(mydat), n)
test = mydat[-train,]
tissue.test = mydat$tissue[-train]
# tissue1 = "brain_amygdala"
# tissue2 = "brain_hippocampus"
mydat$tissue01 = ifelse(mydat$tissue == "brain_amygdala", 1, 0)
boost.best = gbm(tissue01 ~ .-tissue, data = mydat[train,], distribution = "bernoulli",
n.trees = 500, interaction.depth = 4)
summary(boost.best)
mydat$tissue01 <- NULL
#plot(summary(boost.best))
par(mfrow =c(1,1))
#Plot the most relevant genes
sum2<-summary(boost.best);
plot(sum2[1:100,]$rel.inf, xlab = "Most relevant genes")
yhat.boost = predict(boost.best,
newdata = mydat[-train,],
n.trees = 500)
yhat.boost
#Most relevant genes
# ENSG00000118271.5_TTR
# ENSG00000133048.8_CHI3L1
# ENSG00000183395.4_PMCH
#############################################
#RETRAINING CLASSIFICATION TREE WITH 1 GENES
#############################################
par(mfrow =c(1,1))
set.seed(2)
n = round(dim(mydat)[1]*.5)
train = sample(1:nrow(mydat), n)
test = mydat[-train,]
tissue.test = mydat$tissue[-train]
# fix(mydat)
# relevant genes
# ENSG00000118271.5_TTR
# ENSG00000133048.8_CHI3L1
# ENSG00000183395.4_PMCH
# ENSG00000187608.5_ISG15
tree.dataset = tree(tissue ~
ENSG00000133048.8_CHI3L1,
data = mydat, subset = train)
tree.pred = predict(tree.dataset, test, type = "class")
tree.table = table(tree.pred, tissue.test)
tree.table
((tree.table[1] + tree.table[4]) / sum(tree.table))
# 0.5421687 ENSG00000133048.8_CHI3L1
# 0.6144578 ENSG00000118271.5_TTR
# 0.7228916 ENSG00000183395.4_PMCH
plot(tree.dataset);text(tree.dataset, pretty = 0)
tree.cv = cv.tree(tree.dataset)
plot(tree.cv$size, tree.cv$dev, type = "b")
#optimal is 3 so prune back
prune.genes = prune.tree(tree.dataset, best = 2)
plot(prune.genes); text(prune.genes, pretty=0)
prediction.prune = predict(prune.genes, test, type = "class")
tree.table.prune = table(prediction.prune, tissue.test)
tree.table.prune
((tree.table.prune[1] + tree.table.prune[4]) / sum(tree.table.prune))
# 0.6506024 ENSG00000118271.5_TTR
# 0.686747 ENSG00000133048.8_CHI3L1
# 0.7228916 ENSG00000183395.4_PMCH
########################################
#CLASSIFICATION TREE WITHOUT THE 3 GENES
########################################
set.seed(2)
n = round(dim(mydat)[1]*.5)
train = sample(1:nrow(mydat), n)
test = mydat[-train,]
tissue.test = mydat$tissue[-train]
#relevant genes
# ENSG00000118271.5_TTR
# ENSG00000133048.8_CHI3L1
# ENSG00000183395.4_PMCH
# ENSG00000187608.5_ISG15
tree.dataset = tree(tissue ~.
-ENSG00000133048.8_CHI3L1,
data = mydat, subset = train)
tree.pred = predict(tree.dataset, test, type = "class")
tree.table = table(tree.pred, tissue.test)
tree.table
((tree.table[1] + tree.table[4]) / sum(tree.table))
# 0.626506 ENSG00000133048.8_CHI3L1
# 0.7228916 ENSG00000118271.5_TTR
# 0.746988 ENSG00000183395.4_PMCH
plot(tree.dataset);text(tree.dataset, pretty = 0)
tree.cv = cv.tree(tree.dataset)
plot(tree.cv$size, tree.cv$dev, type = "b")
#optimal is 3 so prune back
prune.genes = prune.tree(tree.dataset, best = 3)
plot(prune.genes);text(prune.genes, pretty=0)
prediction.prune = predict(prune.genes, test, type = "class")
tree.table.prune = table(prediction.prune, tissue.test)
tree.table.prune
((tree.table.prune[1] + tree.table.prune[4]) / sum(tree.table.prune)) * 100
# 65.06024 ENSG00000133048.8_CHI3L1
# 68.6747 ENSG00000183395.4_PMCH
# 0.7228916 ENSG00000118271.5_TTR