Discovering knowledge from medical databases using evolutionary algorithms: Learning rules and causal structures for capturing patterns and causality relationships

Abstract

Data mining, referred to as knowledge discovery in databases (KDD), is the nontrivial process of identifying valid, novel and potentially useful patterns in data. Evolutionary algorithms are employed for representing knowledge in rules and causal structures determined by Bayesian networks. Two medical databases are used to learn the rules for representing the patterns of data in addition to the use of Bayesian networks as causality relationship models among the attributes. Advanced evolutionary algorithms such as generic genetic programming, evolutionary programming and genetic algorithms are used to conduct the learning task.