This project is an application of sentiment analysis to detect depression. This is based on 1.7 million tweets collected from Twitter and other sources. The research paper compares various methods and models that are suitable for detecting depression.
Delhi: Can you imagine that depression could be traced to text? Yes, it is possible with Achintya Jha’s machine learning model. Achintya, a class 12 student of Ahalcon International School, has developed a machine learning model that can detect depression from text. He won the Young Achievers Award from India Achievers Forum for the web interface of his research project. They have used a variety of natural language processing techniques to come up with a model that provides an accuracy of around 96%.
This project is an application of sentiment analysis to detect depression. This is based on 1.7 million tweets collected from Twitter and other sources. The research paper compares various methods and models that are suitable for the detection of depression. The project is made accessible through an interactive web-interface. It understands not only raw text but also emoji, hashtags and hyperlinks – the things we see in tweets. The project’s web application provides users with predictable results within 1-2 seconds and is completely free to use. It also has a micro-blogging forum page where people can talk on topics like mental health, exam-stress, education etc. Together these tools are available under the project name Sentimate (short for sentiment-mate, a mental health friend) that provides a way to detect depression and a platform to talk to other people about it. Achintya tested the model with a list of tweets to find out what a person suffering from depression was going through and the model performed very well.
Prior to this, Achintya volunteered for several NGOs like Zenko and Love Heels Cancer, which work to help people suffering from cancer and the many mental health conditions around it. It was clear that there were many people who suffered from depression but never knew about it. There are many surveys on the Internet that can help detect depression, although they are not as accurate and are very lengthy to take. Achintya’s project asks a person about his mood and gives results in a matter of seconds. When he started his project, he aimed to develop a very fast and reliable way to find out with the support of a community of people willing to help. Thus, one of the key features of the web app is the discussion page, where people can talk and help each other.
Apart from his other achievements, Achintya has demonstrated his skills in computer science at a very young age to get internships in big companies like Tech Mahindra and HappyPlus Consulting.
First published:October 6, 2021, 8:18 PM