# Mental Illness Classification on Social Media Texts using Deep Learning and Transfer Learning

**Authors:** Iqra Ameer, Muhammad Arif, Grigori Sidorov, Helena Gòmez-Adorno, Alexander Gelbukh  
**arXiv:** [2207.01012](https://arxiv.org/abs/2207.01012)  
**Date:** 2022-07-03  
**Summary page:** https://www.summarizepaper.com/en/paper/2207.01012v1/

## Key points

- Social media has become a major medium of communication due to social distance restrictions
- Many individuals suffering from mental illnesses are isolated and rely on online platforms for support
- Approximately 450 million people worldwide are affected by mental illnesses
- Artificial Intelligence (AI) methods are being used to assist mental health providers in decision-making
- Researchers have analyzed user data from the Reddit platform to classify common mental illnesses
- Traditional machine learning, deep learning, and transfer learning models were used in the study
- The AI technology aims to automate the detection process and identify individuals who require emergency assistance
- Previous research has focused on analyzing social media text related to mental health issues, particularly on platforms like Twitter and Reddit
- Various approaches, including traditional machine learning and deep learning techniques, have been employed for mental health text analysis.

## Summary

The use of social media as a major medium of communication has become prevalent due to the current social distance restrictions across the world. However, this shift has resulted in millions of individuals suffering from mental illnesses being isolated and unable to receive in-person help. As a result, they have become more reliant on online platforms to express themselves and seek advice on managing their mental disorders. The World Health Organization estimates that approximately 450 million people are affected by mental illnesses such as depression, anxiety, bipolar disorder, ADHD, and PTSD. To address this issue, researchers have turned to Artificial Intelligence (AI) methods to assist mental health providers in decision-making based on patients' authentic information. These AI innovations have shown promising results in various real-world applications, including healthcare. In this study, the researchers analyze unstructured user data from the Reddit platform and classify five common mental illnesses: depression, anxiety, bipolar disorder, ADHD, and PTSD. The study utilizes traditional machine learning, deep learning, and transfer learning multi-class models to detect mental disorders in individuals. By automating the detection process using AI technology, this effort aims to benefit the public health system by identifying individuals who require emergency assistance and informing appropriate authorities. The rest of the paper is organized as follows: Section 2 provides an overview of existing studies on mental illness in literature. Section 3 explains the problem statement and provides insights into the dataset used for analysis. Section 4 details the methodology applied to detect mental disorders using machine learning techniques. Previous research has focused on analyzing social media text related to mental health issues. Initially centered around Twitter text analysis, recent studies have shifted their focus towards platforms like Reddit. Various approaches ranging from traditional machine learning to advanced deep learning techniques have been employed for mental health text analysis. Machine learning classifiers play a crucial role in analyzing healthcare data and extracting hidden patterns. Conventional machine learning approaches often require significant feature engineering for optimal performance but can provide valuable insights for text classification.