Algorithm 4 function: HOFD.

function [$\nu$] = HOFD(X, y) 01: m = 5; where m is the number of fold for cross-validation 02: rmse = zero(1, m) 03: cv = cvpartition (length (y), ’kfold’, m): cross-validation for $k=1,m$ do: 04:    Xtrain = X(cv.train(k),:) 05:    ytrain = y(cv.train(k),:) 06:    Xtest = X(cv.test(k),:) 07:    ytest = y(cv.test(k),:) 08:    call GP(Xtrain, ytrain): where GP denote the Gaussian process regression 09:    return(mdl): return GP model parameters; 10:    predict(’mdl’, Xtest): test GP model 11:    return(ysp): estimated SBP values 12:    $\mathrm{rmse}(\mathrm{k})=\mathrm{sqrt}(\mathrm{mean}({(\mathrm{ysp}-\mathrm{ytest})}^2))$: evaluation criteria endfor 13: $\nu$ = mean(rmse): where $\nu$ denotes least rmse end