Another complication in applying LDA and Fisher's discriminant to real data occurs when the number of observations of each sample does not exceed the number of samples.
The deepness of a voice is not an accurate guide to whether someone is muscular or weedy. Instead, people rely on a subconscious, mysterious 'x-factor' in the voice to work out someone's physique.
Since human memory is fallible, and of course, new members or visitors don' t always know the history, reputation systems are built into web software to track behavior and how the community judges it.