Replication preserving block pattern in r (coding)

I am struggling with a code. I need to do this in the following steps:

1. Randomly permute within a column (i have 6 columns - so within each of the 6 columns) - 100 reps

2. Some rows for some of the columns have NAs. I want to keep the NAs - not permute them, but only permute the non-NA values

3. Calculate for each repetition/each row max consistency (basically the mode ignoring the NAs)

4. Calculate for each iteration/repetition the average of the mode or average of max consistency across the rows. That is for each iteration i will only have 1 value, so for 100 reps i will have 100 values

5. Plot the histograms of the averages from replications, that is frequency plot of 100 values

My data (larger but this is an example)

df = [url removed, login to view](v1 = c("NA", "Low", "High"),

v2 = c("NA", "High","Low"),

v3 = c("NA", "Low", "Low"))

The code i have for now is:

df22<-replicate(100, {apply(df[ sample(nrow(df)),sample( ncol(df)) ], 2, (function(x)) max(table(x)))}) hist(df22)

At a moment, i am getting a new value for each of the rows for each iteration and i do not get the 1 average of across three observations.

Traitement de Données Langage de programmation R

Nº du projet : #16298801

À propos du projet