决策曲线DCA的绘制

梦里梦见梦不见的4972024-08-05 9:03

今天被要求画一下决策曲线，用Python试了半天也没画出来，最后参考了资料还是用R语言绘制了出来。

数据的格式

将四条曲线画在一张表中

R 复制代码

library(rmda)
library(readxl)

# 读取Excel文件
data <- read_excel("C:/Users/Administrator/Desktop/肺动脉数据1000.xlsx")

baseline.model <- decision_curve(y~Qanadli指数 + AD + 血凝块比率+联合,
                                 data = data,
                                 thresholds = seq(0, .4, by = .01),
                                 bootstraps = 10)


set.seed(123)

# 为每个特征创建决策曲线模型
model_qanadli <- decision_curve(y~Qanadli指数,
                                data = data,
                                thresholds = seq(0, 1, by = .01),
                                bootstraps = 10)

model_ad <- decision_curve(y~ AD,
                           data = data,
                           thresholds = seq(0, 1, by = .01),
                           bootstraps = 10)

model_coagulation <- decision_curve(y~血凝块比率,
                                    data = data,
                                    thresholds = seq(0, 1, by = .01),
                                    bootstraps = 10)

model_combined <- decision_curve(y~联合,
                                 data = data,
                                 thresholds = seq(0, 1, by = .01),
                                 bootstraps = 10)


# 绘制四条决策曲线
plot_decision_curve(list(model_qanadli, model_ad, model_coagulation, model_combined),
                    curve.names = c("Qanadli", "AD", "Coagulation", "Combined"),
                    cost.benefit.axis = FALSE,
                    confidence.intervals = FALSE,
                    standardize = FALSE)