Identification strength: make plotting consistent

1. Rely on actually computed standard deviations in bayestopt_ instead of potentially unset estim_params_
2. Remove arbitrary normalizations/omitted normalization in case of division by 0 in normalization
3. Distinguish between 0 identification and division by 0 due to normalization in plots

(cherry picked from commit 7341e21a)
4 jobs for master in 57 minutes and 43 seconds (queued for 38 seconds)