Hybrid Diagnosis Models for Autism Patients Based on Medical and Sociodemographic Features Using Machine Learning and Multicriteria Decision-Making (MCDM) Techniques: An Evaluation and Benchmarking Framework

Table 9

Result of decision matrix.

Alternatives/criteria		Performance evaluation metric criteria
Hybrid diagnosis models		C1	C2	C3	C4	C5	C6	C7
A1	ReF-decision tree	0.37700	0.00001	0.99312	0.98943	0.98943	0.98944	0.98943
A2	ReF-SVM	2.22900	0.91700	0.95052	0.83245	0.83251	0.83266	0.83245
A3	ReF-naive Bayes	0.13400	0.05000	0.96142	0.84415	0.84367	0.84344	0.84415
A4	ReF-KNN	0.68700	0.51000	0.90306	0.76038	0.74901	0.75549	0.76038
A5	ReF-AdaBoost	0.57700	0.28700	0.98897	0.98528	0.98528	0.98530	0.98528
A6	IG-decision tree	0.47200	0.00200	0.99312	0.98943	0.98943	0.98944	0.98943
A7	IG-SVM	2.70600	0.91400	0.94778	0.83358	0.83364	0.83401	0.83358
A8	IG-naive Bayes	0.21900	0.04000	0.96313	0.84113	0.84029	0.83988	0.84113
A9	IG-KNN	0.60900	0.60600	0.89653	0.75585	0.74503	0.74750	0.75585
A10	IG-AdaBoost	0.69700	0.37600	0.98642	0.98189	0.98189	0.98193	0.98189
A11	Chi^2-decision tree	0.46600	0.00200	0.99271	0.98830	0.98830	0.98830	0.98830
A12	Chi^2-SVM	3.19100	1.06000	0.94979	0.83094	0.83080	0.83066	0.83094
A13	Chi^2-naive Bayes	0.22900	0.04300	0.96235	0.84075	0.83966	0.83926	0.84075
A14	Chi^2-KNN	0.90700	0.62900	0.89462	0.74302	0.73201	0.73414	0.74302
A15	Chi^2-AdaBoost	0.75400	0.59500	0.98670	0.98226	0.98226	0.98230	0.98226

C: criteria; A: alternative; C1: train time; C2: test time; C3: AUC; C4: classification accuracy; C5: F1 score; C6: precision; C7: recall.