A team of Korean researchers has developed a system for predicting when suicides are more likely to occur using data available on social networking services.
This is the first time SNS posts have been used to predict suicide rates, according to professor Kim Doh-kwan of Samsung Medical Center, who led the research in collaboration with the data analysis company Daumsoft
“Using suicide forecasting systems could make prevention programs more effective and ultimately contribute to shedding the title of the OECD member with the highest suicide rate,” Kim said.
In 2011, Korea’s suicide rate came in at 31.7 per 100,000 people, topping the list of members of the Organization for Economic Cooperation and Development.
By comparing the suicide rate from the Jan. 1, 2008 to Dec. 31, 2009 period with suicide-related words posted on various social networking services during the period, the team was able to discern a pattern.
The team then combined the data with that regarding economic, social and weather-related factors known to influence suicides to develop a suicide forecast system.
According to the Samsung Medical Center, the system showed 79 percent accuracy when it was used to predict when suicides were more likely in 2010.
According to researchers who took part in the project, the accuracy of the system could be increased to as high as 90 percent if a more data is included in the analysis.
By Choi He-suk (
cheesuk@heraldcorp.com)
