a couple of issues from Florian

[aliaksah]

Concerning Example 8, there is one more problem:

At the very end, when doing predictions with the parallel version, the following gives an error:

bm_parallel <- predict(get.best.model(result_parallel), df[,-1],link = function(x)(1/(1+exp(-x))))
mean(round(bm_parallel)==df$y)

Error in x.precalc %*% object$coefs[object$coefs != 0] :
non-conformable arguments
Can you look into this?

Prediction based on the mpm model works.

mpm_parallel <- predict(get.mpm.model(result = result_parallel,family = "binomial",y = df$y,x=df[,-1]), df[,-1],link = function(x)(1/(1+exp(-x))))
mean(round(mpm_parallel)==df$y)

and check MPM documentation for the value