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

Table 4

AIC values for the feature subsets obtained from the backward elimination process.

Model no.	Removed feature	AIC of the resulting feature subset
0	None	-2532.19
1	WHR	-2533.66
2	DIASTOLIC.BP	-2535.19
3	BMI.IMC	-2536.61
4	HEIGHT	-2537.93
5	HDL	-2538.77