They develop an algorithm capable of identifying unhappy users on social networks

Researchers at the Open University of Catalonia (UOC) have developed an algorithm capable of distinguishing users from social media who are unhappy by analyzing the texts and images they share, a tool they hope will be useful in helping to diagnose possible communication problems. mental health.

Research from the Catalan university has also revealed that Spanish-speaking users are more likely than English-speaking users to mention relationship problems when they feel depressed.

The algorithm, trained on searches on Instagram, Facebook and Twitter, has been based on William Glasser’s theory of choice, according to which there are five basic needs that are at the foundation of all human behavior: survival, power, freedom, belonging and fun.

According to experts, These needs influence which image we choose to upload to our Instagram profile..

“How we show ourselves on social networks can provide useful information about behaviors, personalities, perspectives, motives and needs,” said Mohammad Mahdi Dehshibi, who has coordinated the research in the AI ​​for Human Well-being (AIWELL) group of the Studies Computer, Multimedia and Telecommunications of the UOC.

The researchers have worked for two years on a deep learning model that identifies the five needs described by Glasser, using multimodal data such as images, text, biography or geolocation.

To carry out the study, which is published in the journal ‘IEEE Transactions on Affective Computing’, they analyzed 86 Instagram profiles, published in Spanish and Persian.

Relying on neural networks and databases, the experts trained an algorithm to identify the content of the images and classify the textual content, assigning them different labels proposed by psychologists, who compared the results with a database of more than 30,000 images. captions and comments.

Mahdi Dehshibi, who is also a researcher at the imBody Research Laboratory of the Carlos III University of Madrid and the Unconventional Computing Laboratory of the University of the West of England in Bristol, explains it with an example: “Let’s imagine that a cyclist climbs a mountain and, at the top, they can choose between sharing a selfie or a group image”.

“If you choose the selfie, we perceive the need for power, but if you choose the other, we can conclude that, in addition to having fun, the person looks for a way to satisfy their need to belong,” he clarifies.

Researchers have also found that Spanish-speaking users are more likely than English-speaking users to mention relationship problems when feeling down.

“Studying social media data pertaining to non-English speaking users could help build inclusive and diverse tools and models to address mental health issues in people with diverse cultural or linguistic backgrounds,” they write.

The authors believe that their research can help improve preventive measures, from identifying the problem to improving treatments when a person has been diagnosed with a mental disorder.