Fig. 4 | Scientific Reports

Fig. 4

From: An infection prediction model developed from inpatient data can predict out-of-hospital COVID-19 infections from wearable data when controlled for dataset shift

Fig. 4

Monotonic feature transformation of mean temperature feature. Red, hospital dataset; green, wearable dataset (sleep-only features); blue, transformed wearable dataset (sleep-only features). (A) Data distribution of mean temperature feature: red and green shaded areas describe data distribution from hospital and wearable sleep data respectively. Vertical lines mark the 0-100 percentile values in 5% intervals on the x-axis corresponding to each dataset. (B) Monotonic feature transformation curve (black) where feature values with the same percentile value are mapped between two datasets. Dashed lines mark the 0-100 percentile values in 5% intervals on the x-axis for wearable sleep data (green) and on the y-axis for hospital data (red). (C) Data distribution of mean temperature feature: red, green and blue shaded areas describe data distribution from hospital dataset, wearable sleep dataset and transformed wearable sleep dataset respectively. Vertical lines mark the 0-100 percentile values in 5% intervals on the x-axis corresponding to each dataset; blue vertical lines are overlapped with red vertical lines.

Back to article page