Scribd
Upload
10 views1 page

Stress Detection via NLP and Machine Learning

The document presents a framework for detecting stress during social interactions using Natural Language Processing and Machine Learning applied to social media data. It analyzes user-generated text to categorize emotional states and employs various models, with the CNN + BERT hybrid model achieving the highest accuracy of 96.84%. This approach aims to support mental health professionals by providing an early diagnostic tool for stress detection and insights into user behavior trends.

Uploaded by

anupamams004
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
10 views1 page

Stress Detection via NLP and Machine Learning

The document presents a framework for detecting stress during social interactions using Natural Language Processing and Machine Learning applied to social media data. It analyzes user-generated text to categorize emotional states and employs various models, with the CNN + BERT hybrid model achieving the highest accuracy of 96.84%. This approach aims to support mental health professionals by providing an early diagnostic tool for stress detection and insights into user behavior trends.

Uploaded by

anupamams004
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Abstract

Stress Detection During Social Interactions with Natural Language Processing


and Machine Learning

In today’s digital age, where social interactions are increasingly taking place online,
identifying mental health issues such as stress, anxiety, and depression through
computational techniques has become both essential and feasible. This seminar presents an
advanced and automated framework for stress detection using Natural Language
Processing (NLP), Machine Learning (ML), and Deep Learning (DL) methodologies applied
to social media data. With mental health becoming a growing concern, especially post-
pandemic, early detection through digital footprints offers immense potential for proactive
intervention. The proposed study addresses this by analyzing user-generated text from
platforms like Twitter, categorizing emotional states into five primary classes—Happy,
Angry, Sad, Fear, and Surprise—using datasets created through the Text2Emotion library.
Preprocessing techniques such as stop-word removal, tokenization, and text vectorization
(TF-IDF, Word2Vec, Bag of Words) are implemented to prepare the data. Various ML and DL
models including Naïve Bayes, Support Vector Machine (SVM), K-Nearest Neighbors (KNN),
Convolutional Neural Networks (CNN), and Bidirectional Encoder Representations from
Transformers (BERT) are deployed individually and in combination. Among them, the CNN
+ BERT hybrid model yields the best performance with an accuracy of 96.84% and F1-score
of 95.37%. By integrating computational intelligence with social data, the system
automatically detects stress-related emotional expressions, supporting mental health
professionals with a non-invasive early diagnostic tool. It also offers insights into user
behavior trends, helping identify potential stressors in digital interactions. This study
demonstrates that combining advanced NLP and DL techniques can enable accurate and
efficient detection of psychological stress from unstructured text. The findings support the
development of intelligent mental health monitoring systems that can play a significant role
in early intervention, ultimately contributing to improved well-being.

You might also like