Abstract:This study investigates the use of Large Language Models (LLMs) for improved depression detection from users social media data. Through the use of fine-tuned GPT 3.5 Turbo 1106 and LLaMA2-7B models and a sizable dataset from earlier studies, we were able to identify depressed content in social media posts with a high accuracy of nearly 96.0 percent. The comparative analysis of the obtained results with the relevant studies in the literature shows that the proposed fine-tuned LLMs achieved enhanced performance compared to existing state of the-art systems. This demonstrates the robustness of LLM-based fine-tuned systems to be used as potential depression detection systems. The study describes the approach in depth, including the parameters used and the fine-tuning procedure, and it addresses the important implications of our results for the early diagnosis of depression on several social media platforms.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to improve the accuracy of detecting depression from social media platforms by fine - tuning large language models (LLMs). Specifically, the researchers used two large language models, GPT - 3.5 Turbo 1106 and LLaMA2 - 7B, and utilized a large number of data sets from previous studies to successfully identify depressive content in social media posts with an accuracy rate of 96.0%. Compared with traditional machine learning, deep learning, and natural language processing methods, these fine - tuned large language models show better performance in depression detection. #### Main problems: 1. **Improve the accuracy of depression detection**: Traditional methods such as manual self - reporting and questionnaires are widely used, but they are time - consuming and the results are unreliable. User data on social media provides a rich source of information for the early detection of depression, but traditional methods have difficulty capturing subtle emotional changes in language. 2. **Overcome the limitations of traditional models**: Traditional models (such as ML, DL, NLP) have some limitations when handling depression detection tasks, including a lack of understanding of context, insufficient adaptability and generalization ability, inability to effectively integrate domain - specific knowledge, and poor scalability in real - time analysis. 3. **Utilize the advantages of large language models**: Large language models have more complex natural language processing capabilities and can better understand the semantics and emotions in the text. Especially after fine - tuning, they can more accurately identify language patterns related to depression. #### Solutions: - **Fine - tune large language models**: By fine - tuning GPT - 3.5 Turbo 1106 and LLaMA2 - 7B, the models can better adapt to the specific task of depression detection, thereby improving the accuracy and reliability of the models. - **Improve model performance**: The experimental results show that the fine - tuned models achieve an accuracy rate of 96.0% on the test data, which is significantly better than the existing state - of - the - art systems. - **Provide an early diagnosis tool**: By improving the accuracy of depression detection, these models can serve as potential early diagnosis tools to help timely detect signs of depression on social media platforms, thereby promoting early intervention and treatment. In conclusion, this research solves the limitations of traditional methods in depression detection by fine - tuning large language models and provides a more efficient and reliable solution.

Advancing Depression Detection on Social Media Platforms Through Fine-Tuned Large Language Models

Depression Prediction using Machine Learning Algorithms

Harnessing Large Language Models Over Transformer Models for Detecting Bengali Depressive Social Media Text: A Comprehensive Study

Depression Detection on Social Media with Large Language Models

Depression Classification From Tweets Using Small Deep Transfer Learning Language Models

Novel Transformer Based Contextualized Embedding and Probabilistic Features for Depression Detection From Social Media

Detecting and Measuring Depression on Social Media Using a Machine Learning Approach: Systematic Review

Depression symptoms modelling from social media text: an LLM driven semi-supervised learning approach

Detecting and Measuring Depression on Social Media by Machine Learning Approach: a Review (Preprint)

Linguistic Analysis of Hindi-English Mixed Tweets for Depression Detection

Predicting the language of depression from multivariate twitter data using a feature‐rich hybrid deep learning model

Using Large Language Models to Detect Depression From User-Generated Diary Text Data as a Novel Approach in Digital Mental Health Screening: Instrument Validation Study

Depression Detection On Social Media With Reinforcement Learning

Analysis of Deep Learning Techniques for Early Detection of Depression on Social Media Network - A Comparative Study

Depression Detection Via Harvesting Social Media: A Multimodal Dictionary Learning Solution

Enabling Early Health Care Intervention by Detecting Depression in Users of Web-Based Forums using Language Models: Longitudinal Analysis and Evaluation

Machine Learning for Depression Detection on Web and Social Media

Enhancing Depressive Post Detection in Bangla: A Comparative Study of TF-IDF, BERT and FastText Embeddings

Detecting depression tendency with multimodal features

Depression detection from social network data using machine learning techniques