Skip to main content

Table 5 Confusion matrix of algorithm classification vs. actual diagnosis

From: Using heart rate profiles during sleep as a biomarker of depression

  Actual Diagnosis  
Depression Control
Algorithm    Total  
 Depression 72a 20b 87 PPV: 78.3%
 Control 15c 67d 87 FOR: 18.3%
Total 82 92 174  
  Sensitivity: 82.8% FPR: 23.0%  
FNR: 17.2% Specificity: 77.0%
  1. aTrue positives: Individuals with depression who were correctly classified bFalse positives: Healthy controls who were incorrectly classified as depressedcFalse negatives: Individuals with depression who were incorrectly classified as healthy controls dTrue negatives: Healthy controls who were correctly classified
  2. PPV Positive Predictive Value, FOR False Omission Rate, FNR False Negative Rate, False Positive Rate