New algorithm can detect depressed Twitter users
Artwork: REUTERS
The findings could pave the way for social media platforms to proactively flag mental health concerns with users.
According to the magazine IEEE Trans Transaction on Affective Computingthe algorithm determines the mental state of Twitter users by extracting and analyzing 38 data points.
These data are obtained from users’ public profiles, including the content of their posts, their posting times, and the interactions of other users in the social network.
Study co-author Abdul Sadka, director of the Digital Futures Institute at Brunel University London (UK), said: “We tested the algorithm on two large databases and evaluated the results. our study compared with other depression detection techniques. In all cases, our study outperformed existing techniques in accurately classifying depression.”
Scientists say a large number of people with underlying depression around the world do not seek professional help. The cause includes many factors, in which the patient himself is not aware of his mental condition and also the stigma of society. This leads to “serious delays in diagnosis and treatment”, according to the report The Independent.
“In this study, it is possible to identify depression at an early stage by exploiting online social behaviors,” the scientists said. They used about 80% of the information in each database for algorithmic processing and the remaining data to check its accuracy.
After running through the database, the algorithm looks at 38 differentiating factors – such as a user’s positive and negative word usage, the number of friends and followers they have, and how they use emojis – to estimate the user’s mental and emotional state.
The scientists also say the algorithm could be further developed and used for a number of applications such as sentiment analysis and crime investigation.
However, from this study raises questions about data privacy and the need for users’ consent before their public data is used for analysis.
Dr Huiyu Zhou, another co-author of the study from the University of Leicester (UK), said: “The next stage of this research will be to test its validity in other settings or contexts. On the other hand, the technology obtained from this survey could be further developed for other applications, such as e-commerce, recruitment screening or candidate screening.”
“The proposed algorithm is platform-independent, so it can also be easily extended to other social media systems like Facebook or WhatsApp,” added Dr Zhou.
at Blogtuan.info – Source: tuoitre.vn – Read the original article here
