Skip to main content

Advertisement

Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Table 5 Comparative results of two-stage classification on test dataset

From: Deep learning based low-cost high-accuracy diagnostic framework for dementia using comprehensive neuropsychological assessment profiles

  KLOSCAD-N w/ DNNs Proposed Two-stage Classification MMSE w/ DNNs KLOSCAD-N w/o DNNs MMSE w/o DNNs
Accuracy (%) 92.74 92.90 87.74 86.13 84.84
AUC 0.9790 -a 0.9383 0.9349 0.9143
F1 Score 0.7805 0.7800 0.6667 0.6356 0.6179
Sensitivity 0.9287 0.9343 0.8780 0.8621 0.8736
Specificity 0.9195 0.8966 0.8736 0.8612 0.8443
Likelihood Ratio Plus 11.5425 9.0319 6.9446 6.2092 5.6097
Likelihood Ratio Minus 0.0775 0.0732 0.1396 0.1602 0.1498
Positive Predictive Value 0.5673 0.5064 0.4410 0.4136 0.3892
Negative Predictive Value 0.9913 0.9917 0.9844 0.9821 0.9833
Pre Test Odd 0.1136 0.1136 0.1136 0.1136 0.1136
Post Test Odd 1.3111 1.0259 0.7888 0.7053 0.6372
Post Test Probability 0.5673 0.5064 0.4410 0.4136 0.3892
Costb $111,600 $40,030 $6,200 $111,600 $6,200
  1. aSince each stage provides their own probability, single AUC value can not be calculated
  2. bTotal cost for test dataset including 620 subjects