Exploring emotional patterns in social media through NLP models to unravel mental health insights DOI Creative Commons
Nisha P. Shetty,

Yashraj Singh,

Veeraj Hegde

и другие.

Healthcare Technology Letters, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 1, 2025

Abstract This study aimed to develop an advanced ensemble approach for automated classification of mental health disorders in social media posts. The research question was: can fine‐tuned transformer models (XLNet, RoBERTa, and ELECTRA) with Bayesian hyperparameter optimization improve the accuracy disorder text. Three were on a dataset posts labelled 15 distinct disorders. was employed tuning, optimizing learning rate, number epochs, gradient accumulation steps, weight decay. A voting then implemented combine predictions individual models. proposed achieved highest 0.780, outperforming models: XLNet (0.767), RoBERTa (0.775), ELECTRA (0.755). approach, integrating XLNet, optimization, demonstrated improved classifying from method shows promise enhancing digital potentially aiding early detection intervention strategies. Future work should focus expanding dataset, exploring additional techniques, investigating model's performance across different platforms languages.

Язык: Английский

Exploring emotional patterns in social media through NLP models to unravel mental health insights DOI Creative Commons
Nisha P. Shetty,

Yashraj Singh,

Veeraj Hegde

и другие.

Healthcare Technology Letters, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 1, 2025

Abstract This study aimed to develop an advanced ensemble approach for automated classification of mental health disorders in social media posts. The research question was: can fine‐tuned transformer models (XLNet, RoBERTa, and ELECTRA) with Bayesian hyperparameter optimization improve the accuracy disorder text. Three were on a dataset posts labelled 15 distinct disorders. was employed tuning, optimizing learning rate, number epochs, gradient accumulation steps, weight decay. A voting then implemented combine predictions individual models. proposed achieved highest 0.780, outperforming models: XLNet (0.767), RoBERTa (0.775), ELECTRA (0.755). approach, integrating XLNet, optimization, demonstrated improved classifying from method shows promise enhancing digital potentially aiding early detection intervention strategies. Future work should focus expanding dataset, exploring additional techniques, investigating model's performance across different platforms languages.

Язык: Английский

Процитировано

1