Table 3 Comparison of neural network models trained on summarized audio data and binary vectors indicating the presence or absence of symptoms

From: Generative AI and unstructured audio data for precision public health

Data Type

AUROC

Specificity (@0.80 Sens.)

Sensitivity

Summarized Audio Transcripts

0.823

0.70

0.83

Binary Symptom Vectors

0.769

0.60

0.82

Back to article page