Diabetes Detection Models in Mexican Patients by Combining Machine Learning Algorithms and Feature Selection Techniques for Clinical and Paraclinical Attributes: A Comparative Evaluation

Table 6

Galgo genetic algorithm input parameters. Chromosome size was defined as recommended by Trevino and Falciani [54]. The number of solutions is defined to avoid bias. The number of generations is set to allow most of the models to converge. The minimum performance required is defined by the goal fitness.

Parameter	Value
Classifier	Random forest
Chromosome size	5
Max. solutions	300
Max. generations	200
Goal fitness	0.90