Investigação do modelo de aprendizagem híbrido genético-baseado em instâncias

Abstract

This research work investigates the contribution of genetic algorithms to instance based learning methods, particularly k-NN, IB1 and IB2. The main focus of the investigation is to optimize the performance of three algorithms by means of a genetic algorithm. A computational system that implements each method and its genetic variant is described. Results of experiments in some knowledge domains are presented and analised.