Nearly all standard statistical classification algorithms assume some
knowledge of the distribution of the random variables used to classify.
In particular, the multivariate normal distribution is frequently
assumed and the training set used only to estimate the mean vectors
and covariance matrix of the populations. While some deviation from
normality is tolerated, large deviations usually cause problems. In
particular, multimodal distributions cause even most nonparametric
methods to fail. One of the beauties of neural networks is that they
can typically handle even the most complex distributions. The three-
(and four- ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month, and much more.