As social media platforms such as Facebook, Twitter, Instagram, and others continue to evolve, we’ve seen a rise in depression among users over the years. Citing this, tech companies and researchers have come up with ways to detect and solve the mental health issues of its users. Now, researchers from Europe have developed an advanced algorithm that can detect depression in 9 out of 10 Twitter users. Here are the details.
Researchers Develop Depression-Detection Bot for Twitter
A team of researchers from London’s Brunel University and the University of Leicester has come up with an advanced algorithm that can analyze users’ mental conditions based on their Twitter profiles. The algorithm acquires and analyzes 38 distinct data points from a user’s Twitter profile to detect whether they are going through depression or not.
The researchers developed the new algorithm with two databases. While one contained the Twitter history of numerous Twitter users, the other database included information about their mental health. The team used 80% of the data to teach the bot and 20% of the data to train it.
Coming to the working of the algorithm, the bot initially excludes users with fewer than 5 tweets and runs the remaining profiles through natural language software to check for misspelled words and detect abbreviations. Then, it takes into account 38 specific data points, including the use of positive and negative words, emojis, and other elements to determine the users’ mental state.
Upon testing the new depression-detecting Twitter bot using the Tsinghua Twitter Depression Dataset, the researchers were able to achieve an 88.39% of accuracy, which is pretty impressive. The bot achieved an accuracy of 70.69% on the John Hopkins University CLPsych 2015 dataset.
“Anything that’s above 90% is considered excellent in machine learning. So, 88% for one of the two databases is fantastic”, Prof. Abdul Sadka, the Director of the Institute of Digital Futures at the Brunel University, said. “It’s not 100% accurate, but I don’t think at this level any machine learning solution can achieve 100% reliability. However, the closer you get to the 90% figure, the better,” the Director further added.
The researchers suggest that the new algorithm can be extremely helpful in detecting mental issues in social media users. The bot, as per the team, can be expanded to other platforms like Instagram or WhatsApp, and can also be used in criminal investigations in the future.
“The proposed algorithm is platform-independent, so can also be easily extended to other social media systems such as Facebook or WhatsApp,” Prof. Huiyu Zhou, a Professor of Machine Learning at the University of Leicester, said in a statement. “The next stage of this research will be to examine its validity in different environments or backgrounds, and more importantly, the technology raised from this investigation may be further developed to other applications, such as e-commerce, recruitment examination, or candidacy screening,” he further added.
So, what do you think about the new depression-detecting Twitter bot? Let us know your thoughts on it in the comments below, and stay tuned for more such interesting stories.
