Machine learning can predict rare disasters like pandemics and earthquakes

The computational models are unable to predict the next disaster in advance because there is not enough information to make a prediction.

Researchers from Brown University, Massachusetts Institute of Technology and other institutions say that it does not have to be this way.

The scientists described in a recent study published in Nature Computational Sciences how they combined statistical algorithm — which require less data for accurate and efficient predictions — and a powerful machine-learning technique developed at Brown to train it to predict scenarios and probabilities, and sometimes the timeline, of rare events, despite their lack of historical records.

Source:
https://www.brown.edu/news/2022-12-19/extreme-events