-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathexam.R
132 lines (113 loc) · 4.36 KB
/
exam.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
# Exam
#########################################################################################
# READ FIRST
# 1) set working directory
# 2) load the data
# load("exam_workspace.R")
#########################################################################################
list.of.packages <- c("randomForest", "rfUtilities", "caret", "tree", "rpart")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
if(length(new.packages)) {
install.packages(new.packages)
} else {
print(paste("Packages already installed" ))
}
library(randomForest)
library(caret)
library(rfUtilities)
library(rpart)
library(rpart.plot)
library(tree)
source("helper_functions.R") # Import helper functions
set.seed(1437)
DEBUG <- TRUE
FOLDER_DATA <- "data"
FOLDER_CORRECTED <- "corrected_data"
FOLDER_PEAX <- "peax"
FOLDER_DENSITY <- "density_plots"
FOLDER_UNLABELLED <- "unlabelled_candy_raw"
project_path <- getwd()
load_raw_data(FOLDER_DATA)
assign("peaks_labels", read.table("labels.txt", header = TRUE, sep = "\t"), envir = .GlobalEnv)
generate_density_plot(FOLDER_DENSITY, FOLDER_DATA)
if (DEBUG == TRUE) {
setwd(paste(project_path, FOLDER_CORRECTED, sep = "/"))
temp = list.files(pattern="*_out.csv")
list2env(
lapply(setNames(temp, make.names(gsub("*.csv$", "", temp))),
read.table, header=TRUE), envir = .GlobalEnv
)
} else {
build_peaks(FOLDER_DATA, FOLDER_PEAX, FOLDER_CORRECTED)
}
peaks <- get_peaks("*_out.csv")
peaks.data <- get_peaks_data(peaks)
peaks.n_clust <- get_nclust(peaks.data)
peaks.kmeans <- kmeans(peaks.data, peaks.n_clust, nstart = 1000)
plot(x = peaks.data[,1], y = peaks.data[,2], col=peaks.kmeans$cluster, pch=20,
xlab = "t", ylab = "r", main = paste(peaks.n_clust, "k-means clusters", collapse = " "))
# Build: matrix, training matrix
peaks.labels <- peaks_labels
peaks.matrix <- build_matrix(peaks, peaks.kmeans$cluster, peaks.n_clust)
training <- build_training_matrix(peaks.matrix, peaks_labels)
# SPLIT DATA INTO TRAINING AND TEST DATA (75%)
train_split <- floor(0.75 * nrow(training))
train_ind <- sample(seq_len(nrow(training)), size = train_split)
rf.train <- training[train_ind, ]
rf.test <- training[-train_ind, ]
rf.model <- randomForest(candy ~ ., data = rf.train, ntree = 500)
rf.model
predict(rf.model, rf.test)
reprtree:::plot.getTree(rf.model)
# 5-FOLD CROSS-VALIDATION
rf.num <- 5
fold.trainControl <- trainControl(method = "cv", number = rf.num, classProbs = TRUE, savePredictions = TRUE, allowParallel = TRUE, p = 0.75)
fold.model <- train(candy ~ ., data = training, trControl = fold.trainControl, method = "rf", tuneLength = 20, metric = "Accuracy")
fold.model
plot(fold.model)
# REPORT MEAN, ACCURACY, SENSITIVITY AND SPECIFICITY
predicted <- predict(fold.model, rf.test)
sensitivity(rf.test$candy, predicted)
specificity(rf.test$candy, predicted)
accuracy(rf.test$candy, predicted)$PCC
# 5 PEAKS USING Gini index
gini.peaks <- build_gini_index(fold.model)
#########################################################################################
# DROP THE LEARNING SHIT
# FORMULA
tree.formula <- as.formula(
paste("candy ~ ",
paste(gini.peaks$peaks, collapse = "+"),
sep = "")
)
# APPROACH #2
tree.model2 <- rpart(tree.formula,
data=training,
method="class",
control=rpart.control(minsplit = 2, cp = 0.01),
model = TRUE
)
summary(tree.model2)
rpart.plot(tree.model2, type = 4)
plot_dt()
save_dt()
#########################################################################################
# UNLABELED DATA
load_raw_data(FOLDER_UNLABELLED)
if (DEBUG == TRUE) {
setwd(paste(project_path, "corrected_data", sep = "/"))
temp = list.files(pattern="*_out2.csv")
list2env(
lapply(setNames(temp, make.names(gsub("*.csv$", "", temp))),
read.table, header=TRUE), envir = .GlobalEnv
)
} else {
build_peaks(FOLDER_DATA, FOLDER_PEAX, FOLDER_CORRECTED, "_out2.csv")
}
unlabelled <- get_peaks("*_out2.csv")
unlabelled.data <- get_peaks_data(unlabelled)
unlabelled.n_clust <- peaks.n_clust
unlabelled.kmeans <- kmeans(unlabelled.data, unlabelled.n_clust)
unlabelled.matrix <- build_matrix(unlabelled, unlabelled.kmeans$cluster, unlabelled.n_clust)
ulabelled.the_matrix <- build_training_matrix(unlabelled.matrix)
the_result <- predict_result(tree.model2, ulabelled.the_matrix)