|
1
|
|
# pROC: Tools Receiver operating characteristic (ROC curves) with
|
|
2
|
|
# (partial) area under the curve, confidence intervals and comparison.
|
|
3
|
|
# Copyright (C) 2010-2014 Xavier Robin, Alexandre Hainard, Natacha Turck,
|
|
4
|
|
# Natalia Tiberti, Frédérique Lisacek, Jean-Charles Sanchez
|
|
5
|
|
# and Markus Müller
|
|
6
|
|
#
|
|
7
|
|
# This program is free software: you can redistribute it and/or modify
|
|
8
|
|
# it under the terms of the GNU General Public License as published by
|
|
9
|
|
# the Free Software Foundation, either version 3 of the License, or
|
|
10
|
|
# (at your option) any later version.
|
|
11
|
|
#
|
|
12
|
|
# This program is distributed in the hope that it will be useful,
|
|
13
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
14
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
15
|
|
# GNU General Public License for more details.
|
|
16
|
|
#
|
|
17
|
|
# You should have received a copy of the GNU General Public License
|
|
18
|
|
# along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
19
|
|
|
|
20
|
|
roc <- function(...) {
|
|
21
|
3
|
UseMethod("roc")
|
|
22
|
|
}
|
|
23
|
|
|
|
24
|
|
roc.formula <- function (formula, data, ...) {
|
|
25
|
3
|
data.missing <- missing(data)
|
|
26
|
3
|
roc.data <- roc.utils.extract.formula(formula, data, ...,
|
|
27
|
3
|
data.missing = data.missing,
|
|
28
|
3
|
call = match.call())
|
|
29
|
3
|
response <- roc.data$response
|
|
30
|
3
|
predictors <- roc.data$predictors
|
|
31
|
|
|
|
32
|
3
|
if (length(response) == 0) {
|
|
33
|
3
|
stop("Error in the formula: a response is required in a formula of type response~predictor.")
|
|
34
|
|
}
|
|
35
|
|
|
|
36
|
3
|
if (ncol(predictors) == 1) {
|
|
37
|
3
|
roc <- roc.default(response, predictors[, 1], ...)
|
|
38
|
3
|
roc$call <- match.call()
|
|
39
|
3
|
if (!is.null(roc$smooth))
|
|
40
|
3
|
attr(roc, "roc")$call <- roc$call
|
|
41
|
3
|
return(roc)
|
|
42
|
|
}
|
|
43
|
3
|
else if (ncol(predictors) > 1) {
|
|
44
|
3
|
roclist <- lapply(roc.data$predictor.names, function(predictor, formula, m.data, call, ...) {
|
|
45
|
|
# Get one ROC
|
|
46
|
3
|
roc <- roc.default(response, m.data[[predictor]], ...)
|
|
47
|
|
# Update the call to reflect the parents
|
|
48
|
3
|
formula[3] <- call(predictor) # replace the predictor in formula
|
|
49
|
3
|
call$formula <- formula # Replace modified formula
|
|
50
|
3
|
roc$call <- call
|
|
51
|
3
|
return(roc)
|
|
52
|
3
|
}, formula = formula, m.data = predictors, call = match.call(), ...)
|
|
53
|
|
# Set the list names
|
|
54
|
3
|
names(roclist) <- roc.data$predictor.names
|
|
55
|
3
|
return(roclist)
|
|
56
|
|
}
|
|
57
|
|
else {
|
|
58
|
0
|
stop("Invalid formula:at least 1 predictor is required in a formula of type response~predictor.")
|
|
59
|
|
}
|
|
60
|
|
}
|
|
61
|
|
|
|
62
|
|
roc.data.frame <- function(data, response, predictor,
|
|
63
|
|
ret = c("roc", "coords", "all_coords"),
|
|
64
|
|
...) {
|
|
65
|
3
|
ret <- match.arg(ret)
|
|
66
|
|
|
|
67
|
3
|
if (is.character(substitute(response))) {
|
|
68
|
3
|
response_name <- response
|
|
69
|
|
}
|
|
70
|
|
else {
|
|
71
|
3
|
response_name <- deparse(substitute(response))
|
|
72
|
|
}
|
|
73
|
|
|
|
74
|
3
|
if (is.character(substitute(predictor))) {
|
|
75
|
3
|
predictor_name <- predictor
|
|
76
|
|
}
|
|
77
|
|
else {
|
|
78
|
3
|
predictor_name <- deparse(substitute(predictor))
|
|
79
|
|
}
|
|
80
|
|
|
|
81
|
3
|
if (any(! c(response_name, predictor_name) %in% colnames(data))) {
|
|
82
|
|
# Some column is not in data. This could be a genuine error or the user not aware or NSE and wants to use roc_ instead
|
|
83
|
3
|
warning("This method uses non-standard evaluation (NSE). Did you want to use the `roc_` function instead?")
|
|
84
|
|
}
|
|
85
|
|
|
|
86
|
3
|
r <- roc_(data, response_name, predictor_name, ret = ret, ...)
|
|
87
|
|
|
|
88
|
3
|
if (ret == "roc") {
|
|
89
|
3
|
r$call <- match.call()
|
|
90
|
|
}
|
|
91
|
3
|
return(r)
|
|
92
|
|
}
|
|
93
|
|
|
|
94
|
|
roc_ <- function(data, response, predictor,
|
|
95
|
|
ret = c("roc", "coords", "all_coords"),
|
|
96
|
|
...) {
|
|
97
|
3
|
ret <- match.arg(ret)
|
|
98
|
|
|
|
99
|
|
# Ensure the data contains the columns we need
|
|
100
|
|
# In case of an error we want to show the name of the data. If the function
|
|
101
|
|
# was called from roc.data.frame we want to deparse in that environment instead
|
|
102
|
3
|
if (sys.nframe() > 1 && deparse(sys.calls()[[sys.nframe()-1]][[1]]) == "roc.data.frame") {
|
|
103
|
3
|
data_name <- deparse(substitute(data, parent.frame(n = 1)))
|
|
104
|
|
}
|
|
105
|
|
else {
|
|
106
|
3
|
data_name <- deparse(substitute(data))
|
|
107
|
|
}
|
|
108
|
3
|
if (! response %in% colnames(data)) {
|
|
109
|
3
|
stop(sprintf("Column %s not present in data %s", response, data_name))
|
|
110
|
|
}
|
|
111
|
3
|
if (! predictor %in% colnames(data)) {
|
|
112
|
3
|
stop(sprintf("Column '%s' not present in data %s", predictor, data_name))
|
|
113
|
|
}
|
|
114
|
|
|
|
115
|
3
|
r <- roc(data[[response]], data[[predictor]], ...)
|
|
116
|
|
|
|
117
|
3
|
if (ret == "roc") {
|
|
118
|
3
|
r$call <- match.call()
|
|
119
|
3
|
return(r)
|
|
120
|
|
}
|
|
121
|
3
|
else if (ret == "coords") {
|
|
122
|
3
|
co <- coords(r, x = "all", transpose = FALSE)
|
|
123
|
3
|
rownames(co) <- NULL
|
|
124
|
3
|
return(co)
|
|
125
|
|
}
|
|
126
|
3
|
else if (ret == "all_coords") {
|
|
127
|
3
|
co <- coords(r, x = "all", ret="all", transpose = FALSE)
|
|
128
|
3
|
rownames(co) <- NULL
|
|
129
|
3
|
return(co)
|
|
130
|
|
}
|
|
131
|
|
}
|
|
132
|
|
|
|
133
|
|
roc.default <- function(response, predictor,
|
|
134
|
|
controls, cases,
|
|
135
|
|
density.controls, density.cases,
|
|
136
|
|
# data interpretation
|
|
137
|
|
levels=base::levels(as.factor(response)), # precise the levels of the responses as c("control group", "positive group"). Can be used to ignore some response levels.
|
|
138
|
|
|
|
139
|
|
percent=FALSE, # Must sensitivities, specificities and AUC be reported in percent? Note that if TRUE, and you want a partial area, you must pass it in percent also (partial.area=c(100, 80))
|
|
140
|
|
na.rm=TRUE,
|
|
141
|
|
direction=c("auto", "<", ">"), # direction of the comparison. Auto: automatically define in which group the median is higher and take the good direction to have an AUC >= 0.5
|
|
142
|
|
algorithm=6,
|
|
143
|
|
quiet = FALSE,
|
|
144
|
|
|
|
145
|
|
# what computation must be done
|
|
146
|
|
smooth=FALSE, # call smooth.roc on the current object
|
|
147
|
|
auc=TRUE, # call auc.roc on the current object
|
|
148
|
|
ci=FALSE, # call ci.roc on the current object
|
|
149
|
|
plot=FALSE, # call plot.roc on the current object
|
|
150
|
|
|
|
151
|
|
# disambiguate method/n for ci and smooth
|
|
152
|
|
smooth.method="binormal",
|
|
153
|
|
smooth.n=512,
|
|
154
|
|
ci.method=NULL,
|
|
155
|
|
# capture density for smooth.roc here (do not pass to graphical functions)
|
|
156
|
|
density=NULL,
|
|
157
|
|
|
|
158
|
|
# further arguments passed to plot, auc, ci, smooth.
|
|
159
|
|
...
|
|
160
|
|
|
|
161
|
|
) {
|
|
162
|
|
# Check arguments
|
|
163
|
3
|
direction <- match.arg(direction)
|
|
164
|
|
# Response / Predictor
|
|
165
|
3
|
if (!missing(response) && !is.null(response) && !missing(predictor) && !is.null(predictor)) {
|
|
166
|
|
# Forbid case/controls
|
|
167
|
3
|
if ((!missing(cases) && !is.null(cases)) || (!missing(controls) && !is.null(controls))) {
|
|
168
|
3
|
stop("'cases/controls' argument incompatible with 'response/predictor'.")
|
|
169
|
|
}
|
|
170
|
|
# Forbid density
|
|
171
|
3
|
if ((!missing(density.cases) && !is.null(density.cases)) || (!missing(density.controls) && !is.null(density.controls))) {
|
|
172
|
3
|
stop("'density.*' arguments incompatible with 'response/predictor'.")
|
|
173
|
|
}
|
|
174
|
|
|
|
175
|
3
|
original.predictor <- predictor # store a copy of the original predictor (before converting ordered to numeric and removing NA)
|
|
176
|
3
|
original.response <- response # store a copy of the original predictor (before converting ordered to numeric)
|
|
177
|
|
|
|
178
|
|
# Validate levels
|
|
179
|
3
|
if (missing(levels)) {
|
|
180
|
3
|
if (length(levels) > 2) {
|
|
181
|
3
|
warning("'response' has more than two levels. Consider setting 'levels' explicitly or using 'multiclass.roc' instead")
|
|
182
|
3
|
levels <- levels[1:2]
|
|
183
|
|
}
|
|
184
|
3
|
else if (length(levels) < 2) {
|
|
185
|
0
|
stop("'response' must have two levels")
|
|
186
|
|
}
|
|
187
|
3
|
ifelse(quiet, invisible, message)(sprintf("Setting levels: control = %s, case = %s", levels[1], levels[2]))
|
|
188
|
|
}
|
|
189
|
3
|
else if (length(levels) != 2) {
|
|
190
|
3
|
stop("'levels' argument must have length 2")
|
|
191
|
|
}
|
|
192
|
|
|
|
193
|
|
# ensure predictor is numeric or ordered
|
|
194
|
3
|
if (!is.numeric(predictor)) {
|
|
195
|
3
|
if (is.ordered(predictor)) {
|
|
196
|
3
|
predictor <- tryCatch(
|
|
197
|
|
{
|
|
198
|
3
|
as.numeric(as.character(predictor))
|
|
199
|
|
},
|
|
200
|
3
|
warning = function(warn) {
|
|
201
|
3
|
warning("Ordered predictor converted to numeric vector. Threshold values will not correspond to values in predictor.")
|
|
202
|
3
|
return(as.numeric(predictor))
|
|
203
|
|
}
|
|
204
|
|
)
|
|
205
|
|
}
|
|
206
|
|
else {
|
|
207
|
3
|
stop("Predictor must be numeric or ordered.")
|
|
208
|
|
}
|
|
209
|
|
}
|
|
210
|
3
|
if (is.matrix(predictor)) {
|
|
211
|
3
|
warning("Deprecated use a matrix as predictor. Unexpected results may be produced, please pass a numeric vector.")
|
|
212
|
|
}
|
|
213
|
3
|
if (is.matrix(response)) {
|
|
214
|
3
|
warning("Deprecated use a matrix as response. Unexpected results may be produced, please pass a vector or factor.")
|
|
215
|
|
}
|
|
216
|
|
# also make sure response and predictor are vectors of the same length
|
|
217
|
3
|
if (length(predictor) != length(response)) {
|
|
218
|
3
|
stop("Response and predictor must be vectors of the same length.")
|
|
219
|
|
}
|
|
220
|
|
# remove NAs if requested
|
|
221
|
3
|
if (na.rm) {
|
|
222
|
3
|
nas <- is.na(response) | is.na(predictor)
|
|
223
|
3
|
if (any(nas)) {
|
|
224
|
3
|
na.action <- grep(TRUE, nas)
|
|
225
|
3
|
class(na.action) <- "omit"
|
|
226
|
3
|
response <- response[!nas]
|
|
227
|
3
|
attr(response, "na.action") <- na.action
|
|
228
|
3
|
predictor <- predictor[!nas]
|
|
229
|
3
|
attr(predictor, "na.action") <- na.action
|
|
230
|
|
}
|
|
231
|
|
}
|
|
232
|
3
|
else if(any(is.na(c(predictor[response==levels[1]], predictor[response==levels[2]], response)))) # Unable to compute anything if there is any NA in the response or in the predictor data we want to consider !
|
|
233
|
3
|
return(NA)
|
|
234
|
3
|
splitted <- split(predictor, response)
|
|
235
|
3
|
controls <- splitted[[as.character(levels[1])]]
|
|
236
|
3
|
if (length(controls) == 0)
|
|
237
|
3
|
stop("No control observation.")
|
|
238
|
3
|
cases <- splitted[[as.character(levels[2])]]
|
|
239
|
3
|
if (length(cases) == 0)
|
|
240
|
3
|
stop("No case observation.")
|
|
241
|
|
|
|
242
|
|
# Remove patients not in levels
|
|
243
|
3
|
patients.in.levels <- response %in% levels
|
|
244
|
3
|
if (!all(patients.in.levels)) {
|
|
245
|
3
|
response <- response[patients.in.levels]
|
|
246
|
3
|
predictor <- predictor[patients.in.levels]
|
|
247
|
|
}
|
|
248
|
|
|
|
249
|
|
# Check infinities
|
|
250
|
3
|
if (any(which <- is.infinite(predictor))) {
|
|
251
|
0
|
warning("Infinite values(s) in predictor, cannot build a valid ROC curve. NaN returned instead.")
|
|
252
|
0
|
return(NaN)
|
|
253
|
|
}
|
|
254
|
|
}
|
|
255
|
|
|
|
256
|
|
# Cases / Controls
|
|
257
|
3
|
else if (!missing(cases) && !is.null(cases) && !missing(controls) && !is.null(controls)) {
|
|
258
|
|
# Forbid density
|
|
259
|
3
|
if ((!missing(density.cases) && !is.null(density.cases)) || (!missing(density.controls) && !is.null(density.controls))) {
|
|
260
|
3
|
stop("'density.*' arguments incompatible with 'response/predictor'.")
|
|
261
|
|
}
|
|
262
|
|
# remove nas
|
|
263
|
3
|
if (na.rm) {
|
|
264
|
3
|
if (any(is.na(controls)))
|
|
265
|
3
|
controls <- na.omit(controls)
|
|
266
|
3
|
if (any(is.na(cases)))
|
|
267
|
3
|
cases <- na.omit(cases)
|
|
268
|
|
}
|
|
269
|
3
|
else if (any(is.na(c(controls, cases)))) # Unable to compute anything if there is any NA in the data we want to consider !
|
|
270
|
3
|
return(NA)
|
|
271
|
|
# are there empty cats?
|
|
272
|
3
|
if (length(controls) == 0)
|
|
273
|
3
|
stop("No control observation.")
|
|
274
|
3
|
if (length(cases) == 0)
|
|
275
|
3
|
stop("No case observation.")
|
|
276
|
|
|
|
277
|
|
# check data consistency
|
|
278
|
3
|
if (is.ordered(cases)) {
|
|
279
|
3
|
if (is.ordered(controls)) {
|
|
280
|
3
|
if (identical(attr(cases, "levels"), attr(controls, "levels"))) {
|
|
281
|
|
# merge
|
|
282
|
3
|
original.predictor <- ordered(c(as.character(cases), as.character(controls)), levels = attr(controls, "levels"))
|
|
283
|
|
# Predictor, control and cases must be numeric from now on
|
|
284
|
3
|
predictor <- as.numeric(original.predictor)
|
|
285
|
3
|
controls <- as.numeric(controls)
|
|
286
|
3
|
cases <- as.numeric(cases)
|
|
287
|
|
}
|
|
288
|
|
else {
|
|
289
|
0
|
stop("Levels of cases and controls differ.")
|
|
290
|
|
}
|
|
291
|
|
}
|
|
292
|
|
else {
|
|
293
|
0
|
stop("Cases are of ordered type but controls are not.")
|
|
294
|
|
}
|
|
295
|
|
}
|
|
296
|
3
|
else if (is.numeric(cases)) {
|
|
297
|
3
|
if (is.numeric(controls)) {
|
|
298
|
|
# build response/predictor
|
|
299
|
3
|
predictor <- c(controls, cases)
|
|
300
|
3
|
original.predictor <- predictor
|
|
301
|
|
}
|
|
302
|
|
else {
|
|
303
|
0
|
stop("Cases are of numeric type but controls are not.")
|
|
304
|
|
}
|
|
305
|
|
}
|
|
306
|
|
else {
|
|
307
|
0
|
stop("Cases and controls must be numeric or ordered.")
|
|
308
|
|
}
|
|
309
|
|
|
|
310
|
|
# Check infinities
|
|
311
|
3
|
if (any(which <- is.infinite(predictor))) {
|
|
312
|
3
|
warning("Infinite values(s) in predictor, cannot build a valid ROC curve. NaN returned instead.")
|
|
313
|
3
|
return(NaN)
|
|
314
|
|
}
|
|
315
|
|
|
|
316
|
3
|
response <- c(rep(0, length(controls)), rep(1, length(cases)))
|
|
317
|
3
|
original.response <- response
|
|
318
|
3
|
levels <- c(0, 1)
|
|
319
|
|
}
|
|
320
|
3
|
else if (!missing(density.cases) && !is.null(density.cases) && !missing(density.controls) && !is.null(density.controls)) {
|
|
321
|
3
|
if (!is.numeric(density.cases) || !is.numeric(density.controls))
|
|
322
|
0
|
stop("'density.cases' and 'density.controls' must be numeric values of density (over the y axis).")
|
|
323
|
3
|
if (direction == "auto")
|
|
324
|
3
|
dir <- "<"
|
|
325
|
|
else
|
|
326
|
0
|
dir <- direction
|
|
327
|
3
|
smooth.roc <- smooth.roc.density(density.controls=density.controls, density.cases=density.cases, percent=percent, direction=dir)
|
|
328
|
3
|
class(smooth.roc) <- "smooth.roc"
|
|
329
|
3
|
smooth.roc <- sort(smooth.roc) # sort se and sp
|
|
330
|
|
# anchor SE/SP at 0/100
|
|
331
|
3
|
smooth.roc$specificities <- c(0, as.vector(smooth.roc$specificities), ifelse(percent, 100, 1))
|
|
332
|
3
|
smooth.roc$sensitivities <- c(ifelse(percent, 100, 1), as.vector(smooth.roc$sensitivities), 0)
|
|
333
|
3
|
smooth.roc$percent <- percent # keep some basic roc specifications
|
|
334
|
3
|
smooth.roc$direction <- direction
|
|
335
|
3
|
smooth.roc$call <- match.call()
|
|
336
|
3
|
if (auc) {
|
|
337
|
3
|
smooth.roc$auc <- auc(smooth.roc, ...)
|
|
338
|
3
|
if (direction == "auto" && smooth.roc$auc < roc.utils.min.partial.auc.auc(smooth.roc$auc)) {
|
|
339
|
0
|
smooth.roc <- roc.default(density.controls=density.controls, density.cases=density.cases, levels=levels,
|
|
340
|
0
|
percent=percent, direction=">", auc=auc, ci=ci, plot=plot, ...)
|
|
341
|
0
|
smooth.roc$call <- match.call()
|
|
342
|
0
|
return(smooth.roc)
|
|
343
|
|
}
|
|
344
|
|
}
|
|
345
|
3
|
if (ci)
|
|
346
|
0
|
warning("CI can not be computed with densities.")
|
|
347
|
3
|
if (plot)
|
|
348
|
0
|
plot.roc(smooth.roc, ...)
|
|
349
|
3
|
return(smooth.roc)
|
|
350
|
|
}
|
|
351
|
|
else {
|
|
352
|
0
|
stop("No valid data provided.")
|
|
353
|
|
}
|
|
354
|
|
|
|
355
|
3
|
if (direction == "auto" && median(controls) <= median(cases)) {
|
|
356
|
3
|
direction <- "<"
|
|
357
|
3
|
ifelse(quiet, invisible, message)("Setting direction: controls < cases")
|
|
358
|
|
}
|
|
359
|
3
|
else if (direction == "auto" && median(controls) > median(cases)) {
|
|
360
|
3
|
direction <- ">"
|
|
361
|
3
|
ifelse(quiet, invisible, message)("Setting direction: controls > cases")
|
|
362
|
|
}
|
|
363
|
|
|
|
364
|
|
# smooth with densities, but only density was provided, not density.controls/cases
|
|
365
|
3
|
if (smooth) {
|
|
366
|
3
|
if (missing(density.controls))
|
|
367
|
3
|
density.controls <- density
|
|
368
|
3
|
if (missing(density.cases))
|
|
369
|
3
|
density.cases <- density
|
|
370
|
|
}
|
|
371
|
|
|
|
372
|
|
# Choose algorithm
|
|
373
|
3
|
if (isTRUE(algorithm == 6)) {
|
|
374
|
3
|
if (is.numeric(predictor)) {
|
|
375
|
3
|
algorithm <- 2
|
|
376
|
|
}
|
|
377
|
|
else {
|
|
378
|
0
|
algorithm <- 3
|
|
379
|
|
}
|
|
380
|
|
}
|
|
381
|
3
|
else if (isTRUE(algorithm == 0)) {
|
|
382
|
0
|
load.suggested.package("microbenchmark")
|
|
383
|
0
|
cat("Starting benchmark of algorithms 2 and 3, 10 iterations...\n")
|
|
384
|
0
|
thresholds <- roc.utils.thresholds(c(controls, cases), direction)
|
|
385
|
0
|
benchmark <- microbenchmark::microbenchmark(
|
|
386
|
|
# "1" = roc.utils.perfs.all.safe(thresholds=thresholds, controls=controls, cases=cases, direction=direction),
|
|
387
|
0
|
"2" = roc.utils.perfs.all.fast(thresholds=thresholds, controls=controls, cases=cases, direction=direction),
|
|
388
|
0
|
"3" = rocUtilsPerfsAllC(thresholds=thresholds, controls=controls, cases=cases, direction=direction),
|
|
389
|
0
|
times = 10
|
|
390
|
|
)
|
|
391
|
0
|
print(summary(benchmark))
|
|
392
|
0
|
if (any(is.na(benchmark))) {
|
|
393
|
0
|
warning("Microbenchmark returned NA. Using default algorithm 1.")
|
|
394
|
0
|
algorithm <- 2
|
|
395
|
|
}
|
|
396
|
0
|
algorithm <- as.integer(names(which.min(tapply(benchmark$time, benchmark$expr, sum))))
|
|
397
|
0
|
cat(sprintf("Selecting algorithm %s.\n", algorithm))
|
|
398
|
|
}
|
|
399
|
3
|
else if (isTRUE(algorithm == 5)) {
|
|
400
|
3
|
thresholds <- length(roc.utils.thresholds(c(controls, cases), direction))
|
|
401
|
3
|
if (thresholds > 55) { # critical number determined in inst/extra/algorithms.speed.test.R
|
|
402
|
3
|
algorithm <- 2
|
|
403
|
|
} else {
|
|
404
|
3
|
algorithm <- 3
|
|
405
|
|
}
|
|
406
|
|
}
|
|
407
|
|
|
|
408
|
3
|
if (isTRUE(algorithm == 2)) {
|
|
409
|
3
|
fun.sesp <- roc.utils.perfs.all.fast
|
|
410
|
|
}
|
|
411
|
3
|
else if (isTRUE(algorithm == 3)) {
|
|
412
|
3
|
fun.sesp <- rocUtilsPerfsAllC
|
|
413
|
|
}
|
|
414
|
3
|
else if (isTRUE(algorithm == 1)) {
|
|
415
|
3
|
fun.sesp <- roc.utils.perfs.all.safe
|
|
416
|
|
}
|
|
417
|
3
|
else if (isTRUE(algorithm == 4)) {
|
|
418
|
3
|
fun.sesp <- roc.utils.perfs.all.test
|
|
419
|
|
}
|
|
420
|
|
else {
|
|
421
|
0
|
stop("Unknown algorithm (must be 0, 1, 2, 3, 4 or 5).")
|
|
422
|
|
}
|
|
423
|
|
|
|
424
|
3
|
roc <- roc.cc.nochecks(controls, cases,
|
|
425
|
3
|
percent=percent,
|
|
426
|
3
|
direction=direction,
|
|
427
|
3
|
fun.sesp=fun.sesp,
|
|
428
|
3
|
smooth = smooth, density.cases = density.cases, density.controls = density.controls, smooth.method = smooth.method, smooth.n = smooth.n,
|
|
429
|
3
|
auc, ...)
|
|
430
|
|
|
|
431
|
3
|
roc$call <- match.call()
|
|
432
|
3
|
if (smooth) {
|
|
433
|
3
|
attr(roc, "roc")$call <- roc$call
|
|
434
|
3
|
attr(roc, "roc")$original.predictor <- original.predictor
|
|
435
|
3
|
attr(roc, "roc")$original.response <- original.response
|
|
436
|
3
|
attr(roc, "roc")$predictor <- predictor
|
|
437
|
3
|
attr(roc, "roc")$response <- response
|
|
438
|
3
|
attr(roc, "roc")$levels <- levels
|
|
439
|
|
}
|
|
440
|
3
|
roc$original.predictor <- original.predictor
|
|
441
|
3
|
roc$original.response <- original.response
|
|
442
|
3
|
roc$predictor <- predictor
|
|
443
|
3
|
roc$response <- response
|
|
444
|
3
|
roc$levels <- levels
|
|
445
|
|
|
|
446
|
3
|
if (auc) {
|
|
447
|
3
|
attr(roc$auc, "roc") <- roc
|
|
448
|
|
}
|
|
449
|
|
|
|
450
|
|
# compute CI
|
|
451
|
3
|
if (ci)
|
|
452
|
3
|
roc$ci <- ci(roc, method=ci.method, ...)
|
|
453
|
|
# plot
|
|
454
|
3
|
if (plot)
|
|
455
|
1
|
plot.roc(roc, ...)
|
|
456
|
|
|
|
457
|
|
# return roc
|
|
458
|
3
|
return(roc)
|
|
459
|
|
}
|
|
460
|
|
|
|
461
|
|
#' Creates a ROC object from response, predictor, ... without argument checking. Not to be exposed to the end user
|
|
462
|
|
roc.rp.nochecks <- function(response, predictor, levels, ...) {
|
|
463
|
3
|
splitted <- split(predictor, response)
|
|
464
|
3
|
controls <- splitted[[as.character(levels[1])]]
|
|
465
|
3
|
if (length(controls) == 0)
|
|
466
|
0
|
stop("No control observation.")
|
|
467
|
3
|
cases <- splitted[[as.character(levels[2])]]
|
|
468
|
3
|
if (length(cases) == 0)
|
|
469
|
0
|
stop("No case observation.")
|
|
470
|
3
|
roc.cc.nochecks(controls, cases, ...)
|
|
471
|
|
}
|
|
472
|
|
|
|
473
|
|
#' Creates a ROC object from controls, cases, ... without argument checking. Not to be exposed to the end user
|
|
474
|
|
roc.cc.nochecks <- function(controls, cases, percent, direction, fun.sesp, smooth, smooth.method, smooth.n, auc, ...) {
|
|
475
|
|
# create the roc object
|
|
476
|
3
|
roc <- list()
|
|
477
|
3
|
class(roc) <- "roc"
|
|
478
|
3
|
roc$percent <- percent
|
|
479
|
|
|
|
480
|
|
# compute SE / SP
|
|
481
|
3
|
thresholds <- roc.utils.thresholds(c(controls, cases), direction)
|
|
482
|
3
|
perfs <- fun.sesp(thresholds=thresholds, controls=controls, cases=cases, direction=direction)
|
|
483
|
|
|
|
484
|
3
|
se <- perfs$se
|
|
485
|
3
|
sp <- perfs$sp
|
|
486
|
|
|
|
487
|
3
|
if (percent) {
|
|
488
|
3
|
se <- se*100
|
|
489
|
3
|
sp <- sp*100
|
|
490
|
|
}
|
|
491
|
|
|
|
492
|
|
# store the computations in the roc object
|
|
493
|
3
|
roc$sensitivities <- se
|
|
494
|
3
|
roc$specificities <- sp
|
|
495
|
3
|
roc$thresholds <- thresholds
|
|
496
|
3
|
roc <- sort(roc)
|
|
497
|
3
|
roc$direction <- direction
|
|
498
|
3
|
roc$cases <- cases
|
|
499
|
3
|
roc$controls <- controls
|
|
500
|
3
|
roc$fun.sesp <- fun.sesp
|
|
501
|
|
|
|
502
|
3
|
if (smooth) {
|
|
503
|
3
|
roc <- smooth.roc(roc, method=smooth.method, n=smooth.n, ...)
|
|
504
|
|
}
|
|
505
|
|
# compute AUC
|
|
506
|
3
|
if (auc)
|
|
507
|
3
|
roc$auc <- auc.roc(roc, ...)
|
|
508
|
|
|
|
509
|
3
|
return(roc)
|
|
510
|
|
}
|
xrobin