Fig. 2

Representation of the identifier as presented to a user. The user enters inputs on the left-hand side. The identifier computes class probabilities by calculating the same Neutral Network function as during the training phase, but now with fixed ‘learnt’ parameters. The data used here is from collection HJB17396. Expert analysis had assessed this collection to be Hebeloma circinans from H. sect. Scabrispora. Thus, in this case, the top ‘guess’ of the identifier is correct and this result would be included in the “Top 1” metric