R/kmPlot.R

In Auburngrads/SMRD: Statistical Methods For Reliability Data

kmPlot <-
function (data.ld)
{
    cdfest.out <- cdfest(data.ld)
    xlim <- range(cdfest.out$q[abs(cdfest.out$q) < 1e+09],
                  cdfest.out$p[abs(cdfest.out$p) < 1e+09])
    
    plot.paper(c(0, 400), 
               c(0, 1), 
               xlab = "Days", 
               ylab = "Survival",
               grids = F)
    
    over.interval <- !(cdfest.out$p == cdfest.out$q)
    cdfest.out$q[cdfest.out$q > 10^20] <- 1.1 * xlim[2]
    cdfest.out$p[cdfest.out$p < 0] <- 0
    
    print(cdfest.out)
    
    segments(cdfest.out$p[over.interval], 
             1 - cdfest.out$prob[over.interval],
             cdfest.out$q[over.interval], 
             1 - cdfest.out$prob[over.interval])
    
    the.median <- approx(x = 1 - cdfest.out$prob[over.interval],
                         y = cdfest.out$p[over.interval], 
                         xout = 0.5)$y
    
    L90 <- approx(x = 1 - cdfest.out$prob[over.interval], 
                  y = cdfest.out$p[over.interval],
                  xout = 0.1)$y
    
    cat("The median = ", the.median, "\n")
    cat("L90 = ", L90, "\n")
    
    return(c(the.median = the.median, 
             L90 = L90, 
             number.cases = summary(data.ld)$number.cases))
    
}