Figure 5 | Scientific Reports

Figure 5

From: Comparison of the diagnostic accuracy of resting-state fMRI driven machine learning algorithms in the detection of mild cognitive impairment

Figure 5

Performance metrics of the RF classifier for all five algorithms for our local dataset: A bar chart that shows (a) the mean accuracy, (b) sensitivity, (c) specificity and (d) ROC-AUC scores for each algorithm calculated from the tenfold cross-validation for the RF classifier. The best and results were acquired when we combined the three metrics where multiple algorithms (GA, NSGA-II, GA, PSO) managed to achieve 83% accuracy. RF random forest, ICC intrinsic connectivity, LCOR local correlation, fALFF fractional amplitude of low frequency fluctuations, SFFS sequential floating forward selection, NSGA-II non-dominated sorting genetic algorithm concatenated metrics: The three metrics combined resulting in 396 (3 × 132) regions.

Back to article page