Posted on: 10/08/2003 02:57 PM

A good example of using a probability model to solove a serious problem is the SpamBayes filtering engine, which uses machine learning and Bayesian inference techniques to compute the probability that a given piece of e-mail is spam. This article demonstrates how to develop univariate probability models in PHP; discusses how to fit empirical data distributions to a theoretical probability distribution; and showcases an important tool for all this -- the Probability Distributions Library (PDL).

Printed from Linux Compatible (http://www.linuxcompatible.org/news/story/php_probability_models_for_web_data.html)