Page 131 - DP-2-3

Design+ ML for predicting Alzheimer’s progression

Appendices

Table A1. Classification performance of machine learning models using complete and selected features

Dataset Classification report
type
Complete Simple RF Tuned RF
features Precision Recall F1-score Support Precision Recall F1-score Support
HC 0.95 0.93 0.94 40 HC 0.95 0.93 0.94 40
MCI 0.70 0.89 0.78 18 MCI 0.70 0.89 0.78 18
AD 1.00 0.75 0.86 16 AD 1.00 0.75 0.86 16
Macro-average 0.88 0.85 0.86 74 Macro-average 0.88 0.85 0.86 74
Confusion matrix Confusion matrix
37 3 0 37 3 0
2 16 0 2 16 0
0 4 12 0 4 12
Simple XGBoost Tuned XGBoost
Precision Recall F1-score Support Precision Recall F1-score Support
HC 0.93 0.93 0.93 40 HC 0.90 0.93 0.91 40
MCI 0.71 0.83 0.77 18 MCI 0.70 0.78 0.74 18
AD 0.92 0.75 0.83 16 AD 0.92 0.75 0.83 16
Macro-average 0.85 0.84 0.84 74 Macro-average 0.84 0.82 0.83 74
Confusion matrix Confusion matrix
37 3 0 37 3 0
2 15 1 3 14 1
1 3 12 1 3 12
Selected Simple RF Tuned RF
features Precision Recall F1-score Support Precision Recall F1-score Support
HC 0.93 0.94 0.93 84 HC 0.97 0.93 0.95 84
MCI 0.63 0.70 0.67 27 MCI 0.69 0.89 0.77 27
AD 0.89 0.74 0.81 23 AD 0.95 0.78 0.86 23
Macro-average 0.82 0.79 0.80 134 Macro-average 0.87 0.87 0.86 134
Confusion matrix Confusion matrix
79 5 0 78 6 0
6 19 2 2 24 1
0 6 17 0 5 18
Simple XGBoost Tuned XGBoost
Precision Recall F1-score Support Precision Recall F1-score Support
HC 0.91 0.95 0.93 84 HC 0.97 0.93 0.95 84
MCI 0.63 0.63 0.63 27 MCI 0.68 0.85 0.75 27
AD 0.89 0.74 0.81 23 AD 0.90 0.78 0.84 23
Macro-average 0.81 0.77 0.79 134 Macro-average 0.85 0.85 0.85 134
Confusion matrix Confusion matrix
80 4 0 78 6 0
8 17 2 2 23 2
0 6 17 0 5 18
Abbreviations: AD: Alzheimer’s disease; HC: Healthy control; MCI: Mild cognitive impairment; RF: Random forest.

Volume 2 Issue 3 (2025) 13 doi: 10.36922/DP025270031