python - Is sklearn.manifold.MDS comparable with MATLAB's mdscale called with the metricsstress parameter? -

i recreating project uses multidimensional scaling (mds) visualise data in final stage. specifically, original work uses matlab's mdscale metricsstress parameter according documentation uses 'squared stress, normalized sum of 4th powers of dissimilarities'.

my preferred environment python , implementation of mds i'm aware of sklearn.manifold.mds uses smacof. here stress 'sum of squared distance of disparities , distances constrained points' nothing said normalisation.

my question is: use sklearn implementation in place of mdscale one, results comparable?