Desing of a genetic algorithm to generate knowledge presumptive of the syndrome metabolic

Abstract

Recent studies by the World Health Organization shows that currently exist in the world 220 million people with diabetes and that number is going to double by 2030, of this number approxima- tely 90% of patients have a high probability of having metabolic syndrome, being this syndrome whose diagnosis and treatment protocol is not yet a standard for international institutions who study this syndrome, so it is necessary to use unconventional techniques for its research. One unconven- tional technique and emerging, but is presenting good results in the diagnosis of certain diseases in human medicine is machine learning, however, to the date there is not its application for metabolic syndrome. In this work is introducing a machine learning to create knowledge for the diagnosis of metabolic syndrome, through a proposed genetic algorithm. Knowledge generation is performed so as to optimize the rate of true and false positives, in addition, to introduce a chromo- somal representation of the knowledge and genetic operators of crossover and mutation.