An algorithm developed by the researchers was able to accurately detect 68% of the positive cases up to two days before the onset of symptoms.