CANBERRA, Nov. 21 (Xinhua) -- Scientists from Australia's national science agency have developed a tool that uses artificial intelligence (AI) and social media to detect outbreaks of acute disease events.
The team from Data61, the data science arm of the Commonwealth Scientific and Industrial Research Organisation (CSIRO), were inspired by the thunderstorm asthma event that hit Melbourne without warning in 2016.
Ten people died and more than 8000 were hospitalized by the event on Nov. 21, 2016, which occurred when rye grass pollen was swept across the city by a sudden cool change, sparking asthma-like symptoms in thousands of people.
On the third anniversary of the event on Thursday, Data61 postdoctoral fellow Aditya Joshi said the key challenge in responding to acute disease was detecting them as early as possible.
"The popularity of social media makes it a valuable source of information for epidemic intelligence," Joshi said in a media release.
"We developed a technique that was able to detect the disease outbreak up to nine hours before it was officially reported and before the first news story broke.
"We can draw upon informal sources such as social media data to understand how acute disease events occur, and we can detect when and where an outbreak is likely to occur. This means hospitals and public health agencies can be as prepared as possible."
The team taught the tool to search millions of posts on twitter for terms that could indicate an outbreak such as "breath" and "coughing."
It then uses two fields of AI -- natural language processing (NLP) and statistical time series modelling -- to ensure the posts containing the words were reporting health conditions.
In addition to thunderstorm asthma the tool can also detect outbreaks of Ebola, influenza and the Zika virus, said the CSIRO.
