rms包Predict函数多值预测异常：单值正常多值返回大量行

问题根源 #

你用cph()构建Cox模型时设置了surv=T，这会让模型保存生存函数的相关信息。而Predict.rms()的默认行为是：对每个输入的预测样本，输出其在原始数据所有观测到的事件时间点上的生存概率（或衍生指标），因此36个样本乘以所有事件时间点的数量，就会产生远超预期的46566行结果。

解决方案 #

以下是几种按需选择的解决方式：

1. 获取每个样本的线性预测值（log风险）

指定fun=linear.predictors，直接返回每个样本的单一线性预测结果，共36行：

Predict(cox_model,
    Risk1=c(5,3,2,1.5,1.5,2,3,2.5,4,2,5.5,6,3,3.5,4,5,4.5,3,2,6,3,5,4,1.8,3,3.5,1.5,2.5,3.5,5,6,4,1.5,5,4,2.5),
    Risk2=c(1,1,1,1,0,0,0,1,0,0,0,1,0,0,1,0,1,0,0,1,0,1,0,0,0,1,1,0,1,1,0,1,1,0,0,0),
    Risk3=c(0,0.07,0,0.03,0.01,0.02,0.01,0,0.05,0,0.04,0.03,0.01,0.01,0.01,0,0.11,0.03,0,0.05,0,0,0.02,0.04,0.01,0,0,0.01,0.03,0,0.01,0,0.06,0,0,0.1),
    fun=linear.predictors)

2. 获取每个样本的风险比（相对风险）

用fun=exp将线性预测值转换为风险比，同样返回36行：

Predict(cox_model,
    Risk1=c(5,3,2,1.5,1.5,2,3,2.5,4,2,5.5,6,3,3.5,4,5,4.5,3,2,6,3,5,4,1.8,3,3.5,1.5,2.5,3.5,5,6,4,1.5,5,4,2.5),
    Risk2=c(1,1,1,1,0,0,0,1,0,0,0,1,0,0,1,0,1,0,0,1,0,1,0,0,0,1,1,0,1,1,0,1,1,0,0,0),
    Risk3=c(0,0.07,0,0.03,0.01,0.02,0.01,0,0.05,0,0.04,0.03,0.01,0.01,0.01,0,0.11,0.03,0,0.05,0,0,0.02,0.04,0.01,0,0,0.01,0.03,0,0.01,0,0.06,0,0,0.1),
    fun=exp)

3. 获取指定时间点的生存概率

如果需要特定时间点的生存概率，比如时间=50时的结果，添加time=50参数，每个样本对应一行：

Predict(cox_model,
    Risk1=c(5,3,2,1.5,1.5,2,3,2.5,4,2,5.5,6,3,3.5,4,5,4.5,3,2,6,3,5,4,1.8,3,3.5,1.5,2.5,3.5,5,6,4,1.5,5,4,2.5),
    Risk2=c(1,1,1,1,0,0,0,1,0,0,0,1,0,0,1,0,1,0,0,1,0,1,0,0,0,1,1,0,1,1,0,1,1,0,0,0),
    Risk3=c(0,0.07,0,0.03,0.01,0.02,0.01,0,0.05,0,0.04,0.03,0.01,0.01,0.01,0,0.11,0.03,0,0.05,0,0,0.02,0.04,0.01,0,0,0.01,0.03,0,0.01,0,0.06,0,0,0.1),
    time=50)

4. 构建模型时去掉surv=T（可选）

如果不需要生存函数相关输出，构建模型时删除surv=T参数，此时Predict()默认返回线性预测值：

cox_model <-cph(Surv(Time,Event==1) ~ Risk1 + Risk2 + Risk3, x = T, y = T, data = df)
# 多值预测直接返回36行线性预测值
Predict(cox_model,
    Risk1=c(5,3,2,1.5,1.5,2,3,2.5,4,2,5.5,6,3,3.5,4,5,4.5,3,2,6,3,5,4,1.8,3,3.5,1.5,2.5,3.5,5,6,4,1.5,5,4,2.5),
    Risk2=c(1,1,1,1,0,0,0,1,0,0,0,1,0,0,1,0,1,0,0,1,0,1,0,0,0,1,1,0,1,1,0,1,1,0,0,0),
    Risk3=c(0,0.07,0,0.03,0.01,0.02,0.01,0,0.05,0,0.04,0.03,0.01,0.01,0.01,0,0.11,0.03,0,0.05,0,0,0.02,0.04,0.01,0,0,0.01,0.03,0,0.01,0,0.06,0,0,0.1))

内容的提问来源于stack exchange，提问作者user2110417