r - 使用 R 中的 pROC 使用单个阈值和 0.5 的阈值梯度改变灵敏度和特异性
问题描述
我正在尝试计算多类图像模型的 ROC。但是由于我没有找到多类分类的最佳方法，所以我将其转换为二进制类。我有 31 类图像。使用二进制方法，我试图分别找到每 31 个类的 ROC。
df <- read.xlsx("data.xlsx",sheetName = 1,header = F)
dn <- as.vector(df$X1) # 31 class
model_info <- read.csv("all_new.csv",stringsAsFactors = F) # details of
model output (Actual labels, Model labels, probabablity values)
头部（模型信息）
Actual_labels App_labels X1st
1 tinea cruris and corporis tinea cruris and corporis tinea cruris and corporis
2 tinea cruris and corporis tinea cruris and corporis tinea cruris and corporis
3 tinea cruris and corporis no diagnosis acne vulgaris
4 eczema eczema eczema
5 eczema no diagnosis psoriasis
6 folliculitis impetigo and pyodermas impetigo and pyodermas
X2nd X3rd X.st.. X2nd.. X3rd..
1 psoriasis herpes zoster 0.89 0.05 0.03
2 psoriasis eczema 0.89 0.03 0.02
3 psoriasis molluscum contagiosum 0.29 0.16 0.14
4 tinea cruris and corporis psoriasis 0.62 0.09 0.08
5 melasma tinea cruris and corporis 0.27 0.27 0.25
6 acne vulgaris psoriasis 0.73 0.07 0.03
头(dn)
[1] "acne vulgaris" "alopecia areata" "anogenital warts"
[4] "bullous pemphigoid" "candidiasis" "chicken pox"
app_call 函数基本上根据模型调用是否为真将概率值转换为 0 或 1
app_call <- function(cut_off, category){
labels_thr <- rep(0,nrow(app_res))
ind <- which(model_info$X.st.. >= cut_off) # index of instances
above threshold
true_val <- which(app_res$App.Diagnosis[ind] == category) # index of instances where actual labels are similar to model labels for 1st class out of 31 class.
labels_thr[ind[true_val]] <- 1
return(labels_thr)}
index0 <- grep(pattern = paste0("^",dn[i],"$"),x = model_info$Actual_labels)
actual_labels <- rep(0,nrow(model_info))
if(length(index)>= 1){
actual_labels[index0] <- 1
actual_labels[-index0] <- 0}
app_labels <- app_call(cut_off = 0.5,category = dn[i])
res <- roc(actual_labels,app_labels)
res1 <- roc(actual_labels,model_info$X.st..)
dput(actual_labels)
c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0,
0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0,
0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0)
dput(app_labels)
c(0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0,
0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0,
0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1,
0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0)
dput(model_info$X.st..)
c(0.89, 0.89, 0.29, 0.62, 0.27, 0.73, 0.44, 0.7, 0.42, 0.56,
0.87, 0.19, 0.72, 0.54, 0.37, 0.46, 0.89, 0.89, 0.88, 0.2, 0.46,
0.75, 0.78, 0.66, 0.5, 0.67, 0.17, 0.85, 0.75, 0.58, 0.97, 0.71,
0.8, 0.29, 0.56, 0.44, 0.6, 0.36, 0.38, 0.31, 0.17, 0.35, 0.99,
0.19, 0.48, 0.51, 0.48, 0.92, 0.39, 0.14, 0.44, 0.55, 0.5, 0.43,
0.38, 0.27, 0.37, 0.47, 0.63, 0.4, 0.78, 0.96, 0.67, 0.7, 0.61,
0.44, 0.44, 0.48, 0.47, 0.35, 0.44, 0.34, 0.34, 0.6, 0.24, 0.49,
0.47, 0.35, 0.62, 0.76, 0.67, 0.51, 0.48, 0.72, 0.81, 0.9, 0.89,
0.92, 0.94, 0.2, 0.87, 0.98, 0.99, 0.98, 0.85, 0.89, 0.87, 0.62,
0.19, 0.86, 0.49, 0.5, 0.2, 0.48, 0.51, 0.49, 0.3, 0.11, 0.88,
0.35, 0.81, 0.59, 0.67, 0.31, 0.94, 0.22, 0.4, 0.52, 0.21, 0.31,
0.19, 0.79, 0.64, 0.33, 0.27, 0.19, 0.74, 0.78, 0.41, 0.35, 0.8,
0.39, 0.28, 0.92, 0.37, 0.66, 0.66, 0.44, 0.77, 0.49, 0.27, 0.32,
0.65, 0.21, 0.95, 0.51, 0.16, 0.33, 0.85, 0.26, 0.61, 0.25, 0.57,
0.26, 0.21, 0.45, 0.44, 0.83, 0.31, 0.35, 0.91, 0.34, 0.89, 0.77,
0.56, 0.31, 0.26, 0.84, 0.15, 0.27, 0.47, 0.73, 0.28, 0.44, 0.52,
0.65, 0.72, 0.38, 0.99, 0.46, 0.35, 0.25, 0.42, 0.82, 0.24, 0.56,
0.31, 0.52, 0.91, 0.94, 0.68, 0.21, 0.28, 0.68, 0.67, 0.61, 0.6,
0.43, 0.37, 0.28, 0.28, 0.24, 0.4, 0.92, 0.9, 0.45, 0.96, 0.53,
0.72, 0.26, 0.47, 0.47, 0.97, 0.49, 0.25, 0.51, 0.3, 0.92, 0.43,
0.33, 0.95, 0.25, 0.51, 0.98, 0.23, 0.51, 0.75, 0.84, 0.54, 0.5,
0.54, 0.33, 0.64, 0.29, 0.93, 0.13, 0.27, 0.93, 0.59, 0.27, 0.81,
0.57, 0.59, 0.47, 0.24, 0.53, 0.53, 0.43, 0.24, 0.94, 0.6, 0.7,
0.23, 0.69, 0.95, 0.95, 0.49, 0.73, 0.31, 0.94, 0.15, 0.85, 0.92,
0.34, 0.95, 0.91, 0.36, 0.55, 0.55, 0.29, 0.86, 0.31, 0.48, 0.48,
0.45, 0.5, 0.49, 0.3, 0.33, 0.39, 0.8, 0.42, 0.51, 0.52, 0.66,
0.19, 0.58, 0.94, 0.51, 0.39, 0.84, 0.95, 0.85, 0.72, 0.35, 0.83,
0.5, 0.91, 0.83, 0.61, 0.79, 0.5, 0.87, 0.3, 0.5, 0.53, 0.22,
0.82, 0.74, 0.73, 0.65, 0.88, 0.31, 0.75, 0.74, 0.92, 0.38, 0.47,
0.26, 0.77, 0.78, 0.82, 0.59, 0.59, 0.33, 0.67, 0.31, 0.67, 0.44,
0.77, 0.61, 0.44, 0.77, 0.83, 0.58, 0.6, 0.78, 0.76, 0.47, 0.72,
0.47, 0.29, 0.14, 0.32, 0.17, 0.56, 0.68, 0.3, 0.46, 0.56, 0.68,
0.61, 0.7, 0.23, 0.39, 0.79, 0.38, 0.32, 0.58, 0.46, 0.5, 0.57,
0.93, 0.4, 0.37, 0.75, 0.76, 0.36, 0.84, 0.19, 0.18, 0.94, 0.53,
0.53, 0.24, 0.23, 0.51, 0.53, 0.84, 0.23, 0.44, 0.85, 0.53, 0.23,
0.56, 0.26, 0.38, 0.78, 0.93, 0.65, 0.22, 0.52, 0.35, 0.47, 0.33,
0.31, 0.65, 0.72, 0.46, 0.44, 0.74, 0.92, 0.99, 0.72, 0.41, 0.18,
0.85, 0.89, 0.31, 0.4, 0.98, 0.46, 0.16, 0.58, 0.25, 0.21, 0.32,
0.43, 0.56, 0.34, 0.35, 0.7, 0.43, 0.17, 0.25, 0.33, 0.44, 0.44,
0.58, 0.74, 0.37, 0.68, 0.52, 0.8, 0.96, 0.52, 0.25, 0.81, 0.94,
1, 0.58, 0.42, 0.46, 0.41, 0.18, 0.37, 0.9, 0.54, 0.29, 0.38,
0.38, 0.53, 0.99, 0.57, 0.44, 0.33, 0.45, 0.95, 0.85, 0.75, 0.19,
0.97, 0.27, 0.94, 0.77, 0.79, 0.57, 0.33, 0.98, 0.47, 0.55, 0.27,
0.43, 0.66, 1, 0.62, 0.34, 0.81, 0.4, 0.56, 0.33, 0.25, 0.4,
0.25, 0.91, 0.28, 0.4, 0.73, 0.32, 0.49, 0.37, 0.19, 0.35, 0.29,
0.77, 0.36, 0.31, 0.85, 0.33, 0.61, 0.63, 0.41, 0.98, 0.28, 0.31,
0.91, 0.34, 0.24, 0.82, 0.46, 0.5, 0.39, 0.72, 0.67, 0.51, 0.41,
0.81, 0.74, 0.5, 0.97, 0.65, 0.44, 0.71, 0.35, 0.84, 0.97, 0.42,
0.75, 0.91, 0.61, 0.94, 0.48, 0.42, 0.63, 0.81, 0.83, 0.66, 0.55,
0.61, 0.41, 0.63, 1, 0.63, 0.41, 0.75, 0.27, 0.28, 0.24, 0.55,
0.35, 0.85, 0.97, 0.64, 0.79, 0.92, 0.47, 0.81, 0.23, 0.16, 0.75,
0.12, 0.43, 0.18, 0.69, 0.21, 0.39, 0.19, 0.85, 0.57, 0.97, 0.56,
0.81, 0.13, 0.4, 0.47, 0.95, 0.43, 0.9, 0.67, 0.36, 0.38, 0.83,
0.97, 0.48, 0.93, 0.67, 0.44, 0.34, 0.83, 0.77, 0.39, 0.56, 0.85,
0.55, 0.22, 0.48, 0.46, 0.59, 0.89, 0.99, 0.57, 0.96, 0.97, 0.95,
0.98, 0.24, 0.89, 0.5, 0.94, 0.6, 0.41, 0.71, 0.5, 0.2, 0.96,
0.18, 0.93, 0.92, 0.85, 0.92, 0.82, 0.48, 0.62, 0.53, 0.59, 0.38,
0.8, 0.49, 0.91, 0.58, 0.94, 0.68, 0.15, 0.96, 0.98, 0.89, 0.84,
0.5, 0.88, 0.29, 0.24, 0.31, 0.29, 0.33, 0.49, 0.33, 0.76, 0.54,
0.88, 0.78, 0.26, 0.52, 0.75, 0.97, 0.93, 0.27, 0.69, 0.19, 0.69,
0.2, 0.21, 0.84, 0.31, 0.19, 0.8, 0.6, 0.19, 0.51, 0.98, 0.27,
0.39, 0.77, 0.95, 0.73, 0.28, 0.79, 0.19, 0.98, 0.77, 0.31, 0.84,
0.35, 0.19, 0.26, 0.82, 0.63, 0.38, 0.38, 0.26, 0.63, 0.65, 0.55,
0.88, 0.6, 0.71, 0.85, 0.99, 0.28, 0.42, 0.65, 0.58, 0.97, 0.35,
0.36, 0.32, 0.79, 0.68, 0.39, 0.45, 0.71, 0.98, 0.34, 0.62, 0.24,
0.55, 0.43, 0.95, 0.32, 0.6, 0.63, 0.98, 0.2, 0.31, 0.9, 0.3,
0.32, 0.37, 0.52, 0.64, 0.9, 0.22, 0.31, 0.39, 0.21, 0.93, 0.64,
0.4, 0.96, 0.31, 0.46, 0.86, 0.56, 0.99, 0.83, 0.87, 0.36, 0.59,
0.98, 0.72, 0.21, 0.52, 0.17, 0.21, 0.42, 0.97, 0.34, 0.96, 0.18,
0.63, 0.45, 0.36, 0.31, 0.48, 0.94, 0.86, 0.16, 0.32, 0.97, 0.29,
0.9, 0.38, 0.88, 0.6, 0.17, 0.19, 0.44, 0.98, 0.35, 0.36, 0.2,
0.39, 0.53, 0.35, 0.57, 0.18, 0.26, 0.17, 0.77, 0.51, 1, 0.17,
0.57, 0.48, 0.58, 0.25, 0.32, 0.33, 0.76, 0.16, 0.13, 0.46, 0.44,
0.31, 0.56, 0.46, 0.6, 0.17, 0.36, 0.34, 0.44, 0.43, 0.86, 0.86,
0.44, 0.34, 0.92, 0.32, 0.78, 0.21, 0.46, 0.92, 0.27, 0.98, 0.52,
0.34, 0.27, 0.59, 0.45, 0.58, 0.27, 0.48, 0.21, 0.24, 0.29, 0.89,
0.25, 0.33, 0.96, 0.56, 0.29, 0.97, 0.98, 0.59, 0.28, 0.22, 0.76,
0.91, 0.92, 0.91, 0.94, 0.83, 0.48, 0.53, 0.56, 0.5, 0.75, 0.4,
0.98, 0.6, 0.74, 0.66, 0.97, 0.62, 0.99, 0.39, 0.89, 0.86, 0.66,
0.92, 0.34, 0.99, 0.69, 0.71, 0.8, 0.47, 0.5, 0.83, 0.83, 0.41,
0.72, 0.98, 0.76, 0.65, 0.71, 0.9, 0.9, 1, 0.4, 0.46, 0.35, 0.72,
0.92, 0.74, 0.44, 0.67, 0.97, 0.88, 0.84, 0.71, 0.45, 0.78, 0.9,
0.72, 0.57, 0.68, 0.85, 0.84, 0.46, 0.91, 0.53, 0.96, 0.49, 0.93,
0.49, 0.37, 0.95, 0.47, 0.87, 0.49, 0.58, 0.64, 0.84, 0.8, 0.49,
0.67, 0.75, 0.44, 0.87, 0.71, 0.47, 0.46, 0.83, 0.74, 0.99, 0.86,
0.64, 0.74, 0.43, 0.44, 0.57, 0.89, 0.67, 0.59, 0.89, 0.45, 0.62,
0.81, 0.93, 0.81, 0.98, 0.95, 0.63, 0.64, 0.96, 0.55, 0.49, 0.59,
0.47, 0.42, 0.6, 0.51, 0.4, 0.3, 0.29, 0.45, 0.94, 0.29, 0.33,
0.14, 0.71, 0.41, 0.6, 0.31, 0.95, 0.94, 0.87, 0.8, 0.53, 0.66,
0.71, 0.19, 0.49, 0.97, 0.48, 0.43, 0.38, 0.4, 0.22, 0.38, 0.27,
0.25, 0.45, 0.75, 0.38, 0.23, 0.92, 0.7, 0.68, 0.17, 0.39, 0.65,
0.38, 0.39, 0.21, 0.28, 0.55, 0.89, 0.24, 0.34, 0.92, 0.31, 0.64,
0.86, 0.94, 0.28, 0.43, 0.44, 0.82, 0.23, 0.81, 0.71, 0.53, 0.96,
0.9, 0.55, 0.83, 0.64, 0.51, 0.32, 0.66, 0.45, 0.72, 0.28, 0.34,
0.98, 0.76, 0.52, 0.95, 0.83, 0.47, 0.9, 0.31, 0.23, 0.61, 0.94,
0.61, 0.42, 0.34, 0.55, 0.33, 0.93, 0.24, 0.51, 0.65, 0.17, 0.81,
0.68, 0.51, 0.78, 0.37, 0.37, 0.99, 0.94, 0.64, 0.59, 0.61, 0.9,
0.88, 0.64, 0.49, 0.09, 0.51, NA, 0.86, 0.45, 0.61, 0.24, 0.85,
0.26, 0.29, 0.21, 0.66, 0.26, 0.47, 0.19, 0.99, 0.51, 0.91, 0.37,
0.56, 0.71, 0.47, 0.44, 0.48, 0.52, 0.22, 0.52, 0.29, 0.46, 0.54,
0.94, 0.24, 0.24, 0.47, 0.37, 0.9, 0.79, 0.81, 0.41, 0.38, 0.71,
0.34, 0.46, 0.23, 0.54, 0.43, 0.85, 0.56, 0.26, 0.9, 0.25, 0.3,
0.39, 0.89, 0.38, 0.18, 0.78, 0.37, 0.45, 0.51, 0.8, 0.61, 0.52,
0.84, 0.4, 0.31, 0.28, 0.24, 0.23, 0.43, 0.77, 0.78, 0.95, 0.9,
0.81, 0.15, 0.77, 0.77, 0.87, 0.75, 0.16, 0.49, 0.23, 0.93, 0.45,
0.33, 0.75, 0.32, 0.75, 0.41, 0.24, 0.46, 0.17, 0.41, 0.45, 0.48,
0.15, 0.66, 0.53, 0.75, 0.57, 0.46, 0.78, 0.24, 0.29, 0.95, 0.77,
0.66, 0.94, 0.27, 0.29, 0.58, 0.6, 0.46, 0.58, 0.84, 0.69, 0.47,
0.45, 0.48, 0.35, 0.89, 0.98, 0.93, 0.2, 0.94, 0.91, 0.75, 0.5,
0.44, 0.69, 0.8, 0.76, 0.85, 0.84, 0.72, 0.25, 0.73, 0.26, 0.93,
0.15, 0.33, 0.3, 0.6, 0.24, 0.21, 0.28, 0.51, 0.79, 0.77, 0.85,
0.52, 0.39, 0.68, 0.83, 0.36, 0.15, 0.87, 0.55)
res1 = roc(actual_labels,app_labels)
res2= roc(actual_labels,model_info$X.st..)
实际标签类中它为“1”且概率阈值 (model_info$X.st..) 值大于 0.5 的调用被命名为 app_labels 的“1”，其余全部为零
res1 和 res2 具有不同的灵敏度和特异性值。
解决方案
ROC 曲线显示了当分类器的决策阈值变化时的敏感性和特异性权衡。通常 ROC 曲线函数期望得到预测值和真值作为输入。
这正是您在运行时所做的：
res2= roc(actual_labels,model_info$X.st..)
但是，您app_labels的性质非常不同：您已经在“正确分类”方面进行了合并，这使其更像是一个扁平的列联表，而不是 ROC 函数所期望的“预测”。所以你不能再使用常规的 ROC 函数，需要手动计算灵敏度和特异性。
TP <- sum(app_labels & actual_labels)
TN <- sum(app_labels & !(actual_labels))
FP <- sum(!(app_labels) & !(actual_labels))
FN <- sum(!(app_labels) & actual_labels)
# Sensitivity:
TP / (TP+FN)
# Specificity:
TN / (TN + FP)
推荐阅读
- android - 没有在android中关闭executorService的调用线程就死了
- c++ - Linux 上的 OpenCV + Eclipse - 构建问题
- reactjs - 如何控制传单反应的CircleMarker的z指数？
- c# - 将 FK 约束异常与 EF Core 中的其他数据库异常区分开来？
- twitter-bootstrap - 多项目轮播的响应性问题（已解决）
- javascript - 创建一个支持函数的持久数据钩子
- vbscript - Adobe Illustrator 脚本
- c# - 如何根据其他 2 个属性的值设置 1 个属性的值？
- testing - jest.mock 和 jest.doMock 之间的区别
- scala - 在 uber jar 而不是单个子模块 jar 中构建 sbt 项目