Depression Detection System using Hybrid Deep Learning and Social Media Data

Depression is one of the most common mental health disorders, which is increasingly being expressed throughonline social media platforms. Depression, however, remains untreated and even goes unnoticed. Motivatedby the recent advances in affective computing and social media analysis, this paper proposes a depressiondetection system that combines text, emoji, and image modalities of social media posts using a hybrid machinelearning approach. The proposed system builds upon the recent advances in the sentiment pretraining stageand the depression-specific fine-tuning stage. The system employs a multi lingual DistilBERT-based textencoder, along with an emoji aware text preprocessing approach, and an EfficientNetB4-based facialexpression model for image analysis. The proposed system combines the power of deep representation learningand metrics based classifiers for detecting sentiment, emotion, and visual affect attributes of the social mediaposts of the users. Extensive experiments using publicly available sentiment, emotion, and depression datasetshave been carried out, which prove the effectiveness of the proposed system. The system achieves good results,with weighted F1-scores greater than 0.87 for multi-class sentiment classification and around 0.90 for binarydepression related emotion detection. The results, however, indicate the fea sibility of the proposed system,which is computationally efficient. Index Terms—Depression detection, social media analysis, hybrid machinelearning, sentiment analysis, deep learning, DistilBERT, EfficientNet. Index Terms—Depression detection,social media analytics, hybrid machine learning, sentiment analysis, deep learning, DistilBERT, EfficientNet.