prior <- c(working =0.99,broken =0.01)

 likelihood <- rbind(working = c(good=0.99,bad=0.01),broken =c(good=0.6,bad=0.4))

 data <- c("bad","bad","bad","bad")

 bayes <- function(prior, likelihood, data)

 {

     posterior <- matrix(0, nrow=length(data), ncol=length(prior))

     dimnames(posterior) <- list(data, names(prior))

     initial_prior = prior

     for(i in 1:length(data))

     {

         posterior[i, ] <-

             prior*likelihood[ , data[i]]/

             sum(prior * likelihood[ , data[i]])

         prior <- posterior[i , ]

     }

     return(rbind(initial_prior,posterior))

 }

 bayes(prior,likelihood,data)

 matplot(bayes(prior,likelihood,data),t='b',lty =1,pch =20,col =c(3,2))