Comparación de algoritmos evolutivos para la optimización en la clasificación de la obesidad en escolares

Abstract

The aim of the study is to compare three different evolutionary algorithms: Real Encoding Particle Swarm Optimization (REPSO-C), Incremental Learning with Genetic Algorithms (ILGA) and Decision Tree with Genetic Algorithm (DT-GA) to determine the percent improvement during each iteration of each of the algorithms on a database related to school obesity. For this purpose, the Keel tool was used, which allowed data loading, preparation, processing and analysis of results, both in training and in the tests performed. The precision found by the execution of each of the algorithms, allows to conclude that the model using decision tree and genetic algorithm DT-GA is the best one due to the level of precision that it possesses.