Hey members , in order to get Bootstrapped DEA Scores , i find some MATLAB packages and some STATA code, but it did not works with me, i find them a bit complexed and i could't apply
Please can anyone one offer me a simple MATLAB code or STATA Code to get Bootstrapped DEA Scores for Malmquist ?
thanks alot
Hey Aymen,
at least in Matlab it is rather easy: If X is a nx1 (for the sake of simplicity) array containing your data and you want to take k bootstrapped-samples on the mean of X, use the following syntax:
Y = bootstrp(k,@mean,X);
Where Y is the kx1 array of the bootstrapped means. Replace the function handle @mean by @median for the median or by any other function you want to use (e.g. @(z) std(z)/sqrt(length(z)) will give you the bootstrapped standard error and @(z) mean(z,'omitnan') will give you the mean ignoring missing values marked with NaN).
Hope that helps, Greetings, David
Hey David
thanks for your answer , but i think that the bootstrap technics in DEA is like a black box. what you said let us to get bootstrapped-samples but i want to get bootstrapped-DEA Scores for Malmquist.
so the question is: Doing bootstrapped-samples , then applying DEA to get scores from the bootstrapped-samples is the same as getting a bootstrapped-DEA Scores from an algorithm?
Hey Aymen!
Sorry that I missunderstood your question! I thought you simply wanted to bootstrap your measured scores. As it seems this is not the same as computing the bootstrapped-DEA scores.
Hope someone else can further shed light into this issue!
Greetings, David
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Coding