Skip to main content

Table 1 Comparison of accuracy metrics for All-data and Pre-diagnosis model predictions

From: Instagram photos reveal predictive markers of depression

 

Mitchell et al. μ

All-data μ ( σ )

Pre-diagnosis μ ( σ )

Recall

0.510

0.697 (0.008)

0.318 (0.012)

Specificity

0.813

0.478 (0.012)

0.833 (0.010)

Precision

0.42

0.604 (0.009)

0.541 (0.009)

Negative Predictive Value

0.858

0.579 (0.008)

0.665 (0.006)

F1

0.461

0.647 (0.003)

0.401 (0.008)

  1. General practitioners’ diagnostic accuracy from (Mitchell et al. [24]) is included for comparison. See see Additional file 1 for definitions of accuracy metrics.