As many as 120 Americans commit suicide every day, which translates to about 45,000 deaths per year. Now, a breakthrough project has shown that machine learning may prove to be an effective method in predicting suicide attempts. The groundbreaking findings by researchers at Florida State University and Vanderbilt University Medical Center may help clinicians identify patients at risk of committing suicide up to two years in advance.
Fifty years of unfruitful suicide prediction research has prompted researchers to examine anonymous electronic health records of about two million patients in Tennessee. The researchers then identified more than 3,200 patients who had a history of suicide attempts. The researchers utilized machine learning to evaluate the records and noted that algorithms were able to identify a combination of risk factors that may positively predict future suicide attempts. Machine learning was able to predict suicide attempts up to two years in advance with 80 to 90 percent accuracy, data showed.
“The machine learns the optimal combination of risk factors. What really matters is how this algorithm and these variables interact with one another as a whole. This kind of work lets us apply algorithms that can consider hundreds of data points in someone’s medical record and potentially reduce them to clinically meaningful information,” said lead researcher Jessica Ribeiro.
Information retrieved from machine learning algorithms could be used to develop suicide alert systems. “Just like you get a cardiovascular risk score, you would get a suicide risk score that is informative for clinicians and helps direct them on what steps to take next,” Ribeiro added.
The latest breakthrough in suicide prediction may further establish the highly-controversial precrime concept. Explicitly portrayed in the Tom Cruise film Minority Report, the precrime concept entails the police’s alleged ability to determine preconceived crime and take advanced actions prior to the crime being committed. In the film, precrime was identified by near-comatose mutants with powerful psychic abilities. In the real world, however, big data and sophisticated security systems have already been used by the police to bolster crime detection.
Japanese multinational conglomerate Hitachi launched and tested its crime-prediction software in several unnamed American cities back in 2015. The software, called Hitachi Visualization Predictive Crime Analytics, sifts through various data such as criminal records, social media and weather reports as well as map and transit information to identify important crime-related patterns that may otherwise go unnoticed. The crime prediction software was also created with improved access to video data.
“Digital technologies, like those from Hitachi Data Systems, that provide real-time, aggregate and contextual data, support public safety initiatives that can transform how law enforcement and other first responder agencies locate, mitigate and prevent crimes, and ultimately make our cities safer places,” said International Data Corporation officer Ruthbea Yesner Clark.
Big data has also become a prominent fixture in precrime detection.
Intrado, LexisNexis and Motorola Solutions have also developed a service that readily scans legal, business and social media data to generate information on certain people and circumstances that police officers might encounter when responding to a 911 call. In addition, New York City police have turned to Facebook “friend” lists to resolve gang killings and burglary rings.