According to experts, work is underway to develop anxiety and depression prediction models utilizing artificial intelligence (AI) and Twitter, one of the world’s largest social media platforms, that might detect symptoms of these disorders before clinical diagnosis.
Preliminary findings from the model revealed the possibility of diagnosing the chance of a person acquiring depression based purely on their social media friends and followers, according to researchers from the University of So Paulo (USP) in Brazil.
The results have been published in the journal Language Resources and Evaluation.
Using Natural Language Processing
While there are several studies using natural language processing (NLP) that focus on depression, anxiety, and bipolar illness, the majority of them analyze English texts and do not fit the characteristics of Brazilians, according to the researchers.
The initial stage in this research was to create SetembroBR, a database of information linked to a corpus of 47 million publicly posted Portuguese texts and a network of links between 3,900 Twitter users. Prior to the poll, these users were supposedly diagnosed with or treated for mental health issues. During the COVID-19 epidemic, tweets were gathered.
First, we manually compiled timelines by examining tweets from around 19,000 individuals, roughly the population of a hamlet or small town.
“We then used two datasets, one for users who reported being diagnosed with a mental health problem and another for control purposes that was chosen at random.” “We wanted to distinguish between people with depression and the general population,” said Ivandre Paraboni, the article’s last author and a USP professor.
Mental health problems
The study also collected tweets from friends and followers since persons with mental health problems preferred to follow particular accounts such as discussion forums, influencers, and celebrities who openly disclosed their melancholy.
The second stage, which is still ongoing, has yielded some preliminary results, such as the ability to predict a person’s chance of developing depression based purely on their social media friends and followers, without considering their own posts.
The researchers used deep learning artificial intelligence (AI) to create four text classifiers and word embeddings (context-dependent mathematical representations of relations between words) using models based on bidirectional encoder representations from transformers (BERT), a machine learning algorithm used for NLP.
These models are analogous to a neural network, which learns contexts and meanings by seeing sequential data interactions, such as words in a phrase. The training input was a sample of 200 tweets randomly chosen from each user.
The best among the models in terms of predicting depression using artificial intelligence
The researchers discovered that BERT fared the best among the models in terms of predicting depression and anxiety. They claimed that because the models examined word sequences and complete sentences, they could see that people suffering from depression, for example, tended to write about themselves, using first-person verbs and phrases, as well as topics such as death, crisis, and psychology.
“The signs of depression that can be detected during a doctor’s visit aren’t always the same as the ones that appear on social media,” Paraboni explained.
“For example, the use of the first-person singular pronouns I and me, was very noticeable, which is considered a classic sign of depression in psychology.” We also noticed that sad users frequently used the heart emoji. How AI could help restore voices to people with speech deficits
“This is widely perceived as a symbol of affection and love, but perhaps psychologists haven’t yet classified it as such,” said Paraboni.
The researchers are currently expanding the database, refining their computational techniques, and upgrading the models to see if they can create a tool for future use in screening prospective sufferers of mental health problems and assisting families and friends of young people at risk of depression and anxiety.