Function Predict() does not accept exponential regression in R

Question

Please try to perform a parametric modeling of survival analysis and function predict() is returning an error that does not recognize a variable, but it is there.

Following example:

base=NULL

base$DIAS= c(7,6,8,6,5,5,5,6,6,11,6,4,5,5,5,5,6,4,5,6)

base$DELTA= c(0,1,0,1,0,1,1,0,1,1,0,0,0,1,1,1,1,0,0,0)

base$Protocolo= c(1,0,1,1,0,1,0,1,0,1,0,1,1,1,1,1,1,1,1,1)

base$TIPAC= c(1,0,1,0,0,0,0,1,1,1,0,1,1,1,1,1,0,0,0,0)

base=data.frame(base)

modelo <- survreg(Surv(DIAS,DELTA) ~ Protocolo+TIPAC,data= base, dist = "exponential")

ptempo <- predict(modelo , type = "quantile", newdata= data.frame(1), p=1:99/100,se=TRUE)

Error in Eval(expr, envir, Enclos) : 'Protocol' object not found

Answer 1

In your code, you use base as an argument data, but I imagine you meant modelo. It is not a good practice to use the same variable name for completely different objects as well (data frame and regression). (Modified after AP editing)

Getting back to your problem, the error occurs because you pass a data frame that does not have a column called Protocolo when he calls the predict. You must pass a new data set of the same format as the data used in the regression (with the same variables used in the formula), or leave blank, to perform the prediction with the initial data. The code below works:

base <- data.frame(DIAS = c(7,6,8,6,5,5,5,6,6,11,6,4,5,5,5,5,6,4,5,6),
                     DELTA = c(0,1,0,1,0,1,1,0,1,1,0,0,0,1,1,1,1,0,0,0),
                     Protocolo = c(1,0,1,1,0,1,0,1,0,1,0,1,1,1,1,1,1,1,1,1),
                     TIPAC = c(1,0,1,0,0,0,0,1,1,1,0,1,1,1,1,1,0,0,0,0))

library(survival)
modelo <- survreg(Surv(DIAS, DELTA) ~ Protocolo + TIPAC, data = base, dist = "exponential")

ptempo <- predict(modelo , type = "quantile", p = 1:99/100, se = TRUE)
str(ptempo)
# List of 2
#  $ fit   : num [1:20, 1:99] 0.1285 0.0929 0.1285 0.1228 0.0929 ...
#  $ se.fit: num [1:20, 1:99] 0.0587 0.0552 0.0587 0.0701 0.0552 ...