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
|
#' @useDynLib ModelMetrics
#' @importFrom Rcpp sourceCpp
NULL
#' Test data
#'
#' @name testDF
#' @docType data
NULL
#' @title Multiclass Log Loss
#'
#' @description Calculated the multi-class log loss
#'
#' @param actual A vector of the labels. Can be \code{numeric, character, or factor}
#' @param predicted matrix of predicted values. Can be \code{matrix, data.frame}
#'
#' @export
mlogLoss <- function(actual, predicted){
if(inherits(actual, c('factor', 'character'))){
actual = as.numeric(as.factor(actual))
}
if(inherits(predicted, c('data.frame'))){
predicted = as.matrix(predicted)
}
eps <- 1e-15
predicted = pmax(pmin(predicted, 1 - eps), eps)
mlogLoss_(actual, predicted)
}
#' @title Multiclass Area Under the Curve
#'
#' @description Calculates the area under the curve for a binary classifcation model
#'
#' @param actual A vector of the labels. Can be \code{numeric, character, or factor}
#' @param predicted A data.frame of predicted values. Can be \code{matrix, data.frame}
#'
#'
#' @examples
#' setosa <- glm(I(Species == 'setosa') ~ Sepal.Length, data = iris, family = 'binomial')
#' versicolor <- glm(I(Species == 'versicolor') ~ Sepal.Length, data = iris, family = 'binomial')
#' virginica <- glm(I(Species == 'virginica') ~ Sepal.Length, data = iris, family = 'binomial')
#'
#' Pred <-
#' data.frame(
#' setosa = predict(setosa, type = 'response')
#' ,versicolor = predict(versicolor, type = 'response')
#' ,virginica = predict(virginica, type = 'response')
#' )
#'
#' Predicted = Pred/rowSums(Pred)
#' Actual = iris$Species
#'
#' mauc(Actual, Predicted)
#'
#' @export
mauc <- function(actual, predicted){
actual <- factor(actual)
Data <- data.frame(predicted, actual)
Outcomes <- length(unique(actual))
simpleAUC <- function(x){
# One-vs-all
y1 = levels(Data$actual)[x]
y <- as.numeric(Data[, "actual"] == y1)
prob <- Data[,x]
AUCs <- auc(y, prob)
return(AUCs)
}
AUCs <- sapply(1:Outcomes, simpleAUC)
list(mauc = mean(AUCs), auc = AUCs)
}
#' @title Confusion Matrix
#' @description Create a confusion matrix given a specific cutoff.
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @export
confusionMatrix <- function(actual, predicted, cutoff = .5){
confusionMatrix_(actual, predicted, cutoff)
}
#' @title Positive Predictive Value
#'
#' @description True Positives / (True Positives + False Positives)
#'
#' @aliases precision
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @examples
#' data(testDF)
#' glmModel <- glm(y ~ ., data = testDF, family="binomial")
#' Preds <- predict(glmModel, type = 'response')
#'
#' ppv(testDF$y, Preds, cutoff = 0)
#' precision(testDF$y, Preds, cutoff = 0)
#'
#' @export
ppv <- function(actual, predicted, cutoff = .5){
ppv_(actual, predicted, cutoff)
}
#' @export
precision <- function(actual, predicted, cutoff = .5){
ppv_(actual, predicted, cutoff)
}
#' @title Negative Predictive Value
#'
#' @description True Negatives / (True Negatives + False Negatives)
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @examples
#' data(testDF)
#' glmModel <- glm(y ~ ., data = testDF, family="binomial")
#' Preds <- predict(glmModel, type = 'response')
#'
#' npv(testDF$y, Preds, cutoff = 0)
#'
#' @export
npv <- function(actual, predicted, cutoff = .5){
npv_(actual, predicted, cutoff)
}
#' @title Recall, Sensitivity, tpr
#'
#' @aliases sensitivity tpr
#'
#' @description True Positives / (True Positives + False Negatives)
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @examples
#' data(testDF)
#' glmModel <- glm(y ~ ., data = testDF, family="binomial")
#' Preds <- predict(glmModel, type = 'response')
#'
#' recall(testDF$y, Preds, cutoff = 0)
#' sensitivity(testDF$y, Preds, cutoff = 0)
#' tpr(testDF$y, Preds, cutoff = 0)
#'
#' @export
recall <- function(actual, predicted, cutoff = .5){
recall_(actual, predicted, cutoff)
}
#' @export
sensitivity <- function(actual, predicted, cutoff = .5){
recall_(actual, predicted, cutoff)
}
#' @export
tpr <- function(actual, predicted, cutoff = .5){
recall_(actual, predicted, cutoff)
}
#' @title Specificity, True negative rate
#'
#' @aliases specificity tnr
#'
#' @description True Negatives / (True Negatives + False Positives)
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @examples
#' data(testDF)
#' glmModel <- glm(y ~ ., data = testDF, family="binomial")
#' Preds <- predict(glmModel, type = 'response')
#'
#' tnr(testDF$y, Preds, cutoff = 0)
#' specificity(testDF$y, Preds, cutoff = 0)
#'
#' @export
tnr <- function(actual, predicted, cutoff = .5){
tnr_(actual, predicted, cutoff)
}
#' @export
specificity <- function(actual, predicted, cutoff = .5){
tnr_(actual, predicted, cutoff)
}
#' @title F1 Score
#' @description Calculates the f1 score
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @export
f1Score <- function(actual, predicted, cutoff = .5){
f1Score_(actual, predicted, cutoff)
}
#' @title F Score
#' @description Calculates the F score and allows different specifications of the beta value (F0.5)
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#' @param beta the desired beta value (lower increases weight of precision over recall). Defaults to 1
#'
#' @export
fScore <- function(actual, predicted, cutoff = .5, beta = 1){
fScore_(actual, predicted, cutoff, beta)
}
#' @title Matthews Correlation Coefficient
#' @description Calculates the Matthews Correlation Coefficient
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @export
mcc <- function(actual, predicted, cutoff){
mcc_(actual, predicted, cutoff)
}
#' @title kappa statistic
#'
#' @description Calculates kappa statistic. Currently build to handle binary values in \code{actual} vector.
#'
#' @param actual A vector of the labels
#' @param predicted A vector of predicted values
#' @param cutoff A cutoff for the predicted values
#'
#' @export
kappa <- function(actual, predicted, cutoff = .5){
kappa_(actual, predicted, cutoff)
}
|
