Desktop Survival Guide
by Graham Williams
With the rather unruly explosion of interest in data mining and its well documented commercial successes, and the fact that data mining is the fusion of many disciplines, each with their own heritage, the terminology used in the data mining community is at times quite confusing and often redundant. The beginnings of a Glossary began here but has ceased. http://en.wikipedia.orgWikipedia is now the canonical source.
Bias: The error in a model due to systematic inadequacies in the learning process. That is, those instances consistently incorrectly classified by models built by the learning algorithm. Modelling error due to bias can be reduced using http://en.wikipedia.org/wiki/BoostingBoosting. Compare with http://en.wikipedia.org/wiki/variancevariance.