Teens voluntarily documenting their personal lives with the help of photos on Instagram became unwitting participants in an experiment to study alcohol habits. Researchers at the University of Rochester have found it much easier to research alcohol addiction, brand popularity and other details, using learning algorithms and social media. Using neural networks and computer image recognition technologies, researchers extract information from both photographs and signatures to them. Although Instagram does not filter by age, using modern algorithms for face recognition, you can accurately determine the age and gender of the account holder from the photos in the profile.
Classical methods of studying the consumption of various foods suffer from low accuracy, since people, especially teenagers, do not always answer questions honestly and in detail. Especially when it comes to alcohol. The researchers are confident that with the help of computer technology it will be possible to build a more complete and accurate picture of alcohol consumption by minors. Scientists will present their work with a detailed description of the technologies at a conference of the Institute of Electrical and Electronics Engineers ( IEEE ) on big data processing.
Among other things, analysis of photographs, captions and tags showed that alcohol consumption patterns in adolescents do not differ much from adults – the amount of alcohol consumed increases on weekends, holidays and towards the end of the day. There were no significant differences in frequency and volume in consumption patterns depending on gender, but it was found that gender preferences exist for different brands of alcohol. Naturally, the main goal of such studies is to reduce alcohol consumption by minors. “We can try several approaches to this, armed with this information,” says Jiebo Luo, professor of computer science at the university and lead author of the paper. “Government services and schools can be better informed and, on this basis, develop more effective alcohol control strategies. We can establish interaction with social networks, so that they give targeted anti-advertising of alcohol, and track the results of such companies. ” However, the researchers also admit that alcohol producers can happily use this information to create targeted advertising for their products.