Written By: Prerna Mittra | Updated : December 25, 2023 5:01 PM IST
Image credits by: Photo: Freepik
For many people, thinking about the future is a scary thought -- more so, if it involves death. Most people do not wish to know exactly how they would kick the bucket. But, new research seems to suggest that one can get a picture of when they might die. According to a CNN report, researchers in Denmark claim to have used "powerful machine-learning algorithms" to "accurately predict" many aspects of life, including how early someone is likely to die.
Published recently in the journal in Nature Computational Science, the study reportedly details how a machine-learning algorithm model called 'life2vec' can predict the outcome of a person's life when presented with "highly specific" data about them. The study's lead author and professor at the Technical University of Denmark Sune Lehmann was quoted as saying that with that data, they can "make any kind of prediction".
The researchers, however, state that since it is a "research prototype", it cannot perform any "real-world tasks" in its current state. Per the CNN report, data was collected from a national register in Denmark, comprising "a diverse grouping of 6 million people". The information from 2008 to 2015 included major aspects of a person's life such -- their education, health, income, occupation, etc. Researchers purportedly adapted 'language processing techniques' to generate a vocabulary so that life2vec could interpret sentences about life events based on the data.
Lehmann added that since the algorithm had learned from that data, it was able to make predictions about certain aspects of lives, including how a person may behave, feel and think, and also whether they were going to die in the next few years.
The team of researchers used data from January 1, 2008 to December 31, 2015 on a cohort of over 2.3 million people between the age bracket of 35 years and 65 years; mortality in this age group is difficult to predict, said Lehmann. Then, life2vec worked on the probability of a person surviving the four years after 2016, based on the data it had received.
"To test how good [life2vec] is, we pick a group of 1,00,000 individuals where half survive and half die," Lehmann was quoted as saying. The researchers knew which people had died after 2016, but the algorithm did not. When they put it to the test, the results were "impressive": the algorithm was correct 78 per cent of the time.
Per the report, researchers found people who were male were more likely to die after 2016. Being a skilled worker, such as an engineer, or having a mental health diagnosislike depression or anxiety also led to an earlier death. Being in a managerial position or having a high income, meanwhile, steered people towards the 'survive' column.
Lehmann said while the algorithm is "ominous and crazy", it is "actually something that there's been a really a lot of work on, especially driven by insurance companies". But, the algorithm's limitations include implementing it in a country where the healthcare system is not as robust, or using it on people who are non-salaried and do not engage with the healthcare systems.
Disclaimer: The content on TheHealthSite.com is only for informational purposes. It is not at all professional medical advice. Always consult your doctor or a healthcare specialist for any questions regarding your health or a medical condition.
