Fig. 4

The mean success rate of obtaining \(r = 1\) for the correct parent formula using the scoring functions \(s_{\text {ne}}\), \(s_{\text {LBJ}}\), and \(s_{\text {v}}\) as a function of the number of likely parent candidates formulae \(N_{\text {cand}}\) (left) and mean RRP (right), for the CASMI data using the alphabets CHNO+ (top), CHNOP+ (middle) and CHNOPF+ (bottom). The number of spectra in each bin is given in parentheses