Monday , January 18 2021

Apply artificial intelligence to thousands of photos of Instagram to predict influenza outbreak



The combination of scientific knowledge and technological applications gives way to a new series of previous unsuspected techniques or treatments in medicine. Two Finnish researchers published a study of a system they used artificial intelligence and images taken by Instagram, which would predict the peak of the influenza epidemic one or two weeks in advance, and that would achieve a precision of over 90%.

The study is based on taking more than 22 thousand people with influenza in Finland over the last five years have posted publications on the social network Instagram,

Finnish scientists specializing in machine learning and artificial intelligence, Mike Ermes and Ouguzan Gencoglu, publish their work in the North American University of Cornell magazine, which is to compare the different kinds of hats (flute, fever, headache, and headache) with photos posted on Instagram of medical boxes, tea cups with lemons and thermometers.

Once the information was introduced, it turned out at the height of the flu outbreak the degree of matching achieved by the algorithm reached 96%, one week before 90%, and 86% two weeks.

When developing the artificial intelligence algorithm to check if there is a relationship between the two databases, one of the texts and the other with images, found that people who still feel bad continue to publish on social networks, in the case of Instagram.

The European country has the long-standing tradition of the health system a strong technical database.


Source link