Mental Health Assessment Based on Natural Language Processing

Chenyu Liu

2024

Abstract

As mental health issues gradually become one of the top global concerns, an effective and scalable assessment tool is needed. Traditional psychological self-assessment scales, such as the Beck depression scale and the self-rating anxiety scale, are widely used in clinical diagnosis. However, they rely on self-report from subjects and are easily influenced by subjective emotions, environmental factors, and comprehension abilities. As a result, they may not always accurately reflect an individual's true psychological state. In addition, the frequency of regular testing is often limited and cannot dynamically track individual emotional fluctuations, especially in the short term. This may lead to missed opportunities for early intervention. The development of Natural Language Processing (NLP) technology has made it possible to analyze potential psychological problems in social media and intervene in advance. This paper proposes an NLP-based framework to detect depression, anxiety, and suicidality from user-generated text. This work uses fine-tuned Bidirectional Encoder Representations from Transformers (BERT) models to classify each mental health state from different dimensions, then employs an ensemble method to detect a person's mental state comprehensively. The system is designed to provide early identification of mental health risks. Experimental results validate the approach's accuracy and its potential impact on mental health interventions.

Paper Citation

in Harvard Style

Liu C. (2024). Mental Health Assessment Based on Natural Language Processing. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM; ISBN 978-989-758-738-2, SciTePress, pages 224-228. DOI: 10.5220/0013281600004558

in Bibtex Style

@conference{mlscm24,
author={Chenyu Liu},
title={Mental Health Assessment Based on Natural Language Processing},
booktitle={Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM},
year={2024},
pages={224-228},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013281600004558},
isbn={978-989-758-738-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM
TI - Mental Health Assessment Based on Natural Language Processing
SN - 978-989-758-738-2
AU - Liu C.
PY - 2024
SP - 224
EP - 228
DO - 10.5220/0013281600004558
PB - SciTePress