Algorithm that distinguishes if social network users are unhappy

source

Several researchers who work at The Open University of Catalonia (UOC), gave themselves the task of developing an algorithm that is able to distinguish if users who use social networks are unhappy and to achieve it, we proceed to perform an analysis of images and texts that share in their profiles.

According to researchers, this algorithmic tool will also help diagnose mental health problems or failing, detect if the person is at risk of suffering from any disorder. During the development of research, they managed to detect that non-English speaking users; They are more likely to speak publicly on social networks about their depression problems and, they made this comparison with English -speaking people.

The algorithm that Instagram, Twitter and Facebook is designed under the theory of choice that Williams Glasser created and is based on the 5 basic needs of behavior that human beings have. These are: freedom, survival, belonging, power and fun. According to the researchers who worked on this new design, at the time of selecting which photography we will publish in our profiles, the aforementioned needs influence 100%.

Mohammad Mahdi Dehshibi, coordinator of this research explained to the group to the for Human Well-Being (Aiwell) for its acronym in English that, the way in which we show ourselves on digital platforms, we begin the process of providing information regarding our personality, Needs, behaviors, motives and the perspective of life we ​​have.

For the development of this new algorithm, the researchers worked for 2 years and as a main source, made use of a learning model that was able to identify the 5 needs raised by Glasser. During the 2 years of studies, they used multimodal data, geolocation, biography and text data. It should be noted that, the 86 Instagram profiles they used during the analysis were from Spanish and Persian speech people.

It is important to note that the researchers relied on databases and neuronal in order to train the algorithm to identify and classify images of images and texts. During this analysis, the algorithmic tool provides psychological behavior labels and this allows the result to be more efficient.

Finally, researchers ensure that this algorithm is a tool that can be used to start preventing mental health problems, during the remaining months of the year, this algorithm will improve and thus, starting 2023 may be anchored to which Use Instagram, Twitter and Facebook.