I have two dataframes i.e. ref and observed.
I would like to extract ordinate and abscissa for each observed 's rain value from the cumulative distribution function (CDF) of each month of ref's rain_total.
id1 <- "1JvSG10RmHbxamhvvWEVyb0AzK_9bC_2m" id2="1Is36zNhvSSRYJA33sG6qj_C6bha6ogEe" ref=read_csv(paste0("https://docs.google.com/uc?id=",id1, "&export=download"),col_names = TRUE) ref=ref %>% filter(month==01) head(ref) observed=read_csv(paste0("https://docs.google.com/uc?id=",id2, "&export=download"),col_names = TRUE) head(observed,10)
I used the following code for CDF using 1st month
cdf <- as.data.frame(Ecdf(ref$rain_total)) cdf plot(cdf, type="l")
Trial for ordinate and abscissa extraction:
#quantile at 20%, 50% and 80% quantile(ecdf(ref$rain_total), c(0.2, 0.5, 0.8)) # ordinate for value 2.54,15,45 predict(loess(cdf$y ~ cdf$x), c(2.54, 15, 45))
This solution does provide a value higher than the expected value like for ordinate at 45 should not >1 and quantile 50% should be <5.08 which is not a good condition in mine case.
I will be grateful if I can get another approached like extracting from the line plot instead of the step function which I have tried.