How the pessimism works

However, in binary (henceforth: in multi-label) TC, accuracy is not an adequate measure. The reason for this is that in binary TC applications the two categories c i and c i' are usually unbalanced, i.e. one contains far more members than the other 4 . In this case, building a classifier that has high accuracy is trivial, since the trivial rejector, i.e. the classifier that trivially assigns all documents to the most heavily populated category (i.e. c i), has indeed very high accuracy;

Fabrizio Sebastiani
Dipartimento di Matematica Pura e Applicata
Universit`a di Padova
35131 Padova, Italy

..very interesting, what implies from it is that pessimism is a privmitive attempt to classify the reality, which is favoured only because it brings satisfatory results...

I never thought of it.. but in the end Murphy's laws are based on it :) wow, that's so obvious ..