This is based on a comment from @mvinaixa: we should make sure we're not calculating the similarity score only on the peaks that match between the compared spectra, but consider also the non-matching peaks because otherwise we overestimate the similarity. I'll have to dig into the functions to see if this is the case (I guess it is - just to be sure).
This is based on a comment from @mvinaixa: we should make sure we're not calculating the similarity score only on the peaks that match between the compared spectra, but consider also the non-matching peaks because otherwise we overestimate the similarity. I'll have to dig into the functions to see if this is the case (I guess it is - just to be sure).