Traditionally there have been two paradigms of statistical analysis — classical and Bayesian. Machine learning is essentially a third paradigm, based on algorithms that rely heavily on the speed of modern computing to derive “decision rules” that predict customer behavior. We discuss several machine learning techniques, including covering algorithms, instance-based learning, genetic algorithms, Bayesian networks, support vector machines, and committee machine methods such as bagging and boosting.
KeywordsSupport Vector Machine Bayesian Network Training Instance Service Recovery Rule Induction
Unable to display preview. Download preview PDF.