This document discusses various computational intelligence methods for predicting customer churn in telecommunication companies. It begins by introducing the problem of high customer churn rates in a competitive telecom market. It then discusses approaches like basic classifiers, data preprocessing techniques, and ensembles of classifiers. The document evaluates several specific techniques - multilayer perceptrons, genetic programming, self-organizing maps, and negative correlation learning. It concludes by discussing future work areas and published research applying these methods to improve churn prediction.