Assessing the Psychological Impact of Internet Blackouts:

A Stress Analysis Case Study

Access to the internet is crucial for communication, information, and essential services in modern life.
Internet blackouts during politically sensitive periods can severely affect mental health, increasing stress,
anxiety, and emotional distress. This study examines the psychological impact of the internet shutdown
during the Bangladesh Quota Movement in July 2024. A survey of 2,085 participants captured
behavioral, emotional, and psychological responses in academic, work, and social contexts. Textual
responses were processed using NLP techniques to extract relevant features reflecting stress levels. An
MLP (Multi-Layer Perceptron) classifier was applied to predict stress intensity across participants. The
model demonstrated high performance in classifying stress, achieving significant accuracy and reliability.
Findings highlight the widespread mental distress caused by internet disruptions. The study emphasizes
the importance of timely mental health interventions during crises. These insights align with SDG 3,
promoting mental well-being and social resilience in low- and middle-income countries.
 EXSISTING SYSTEM PROPOSED SYSTEM
 Support Vector Machine (SVM) is a  The proposed system aims to predict stress levels
powerful supervised learning algorithm during internet shutdowns by analyzing textual
widely used for classification tasks, responses collected from participants. Initially,
including stress level analysis in survey data containing behavioral, emotional, and
psychological studies like ours. In the psychological indicators are preprocessed using
context of analyzing the psychological NLP techniques. This involves text cleaning,
impact of internet blackouts, SVM tokenization, stop-word removal, and
effectively separates different stress vectorization to convert textual responses into
categories by finding the optimal boundary numerical features suitable for machine learning.
between data points representing various The processed features capture nuances in
emotional and behavioral states. Its ability language that reflect stress, anxiety, and
to handle high-dimensional and complex emotional distress levels. By leveraging these
datasets makes it well-suited for features, the system prepares a comprehensive
interpreting diverse survey responses, dataset representing individual responses across
ensuring robust and reliable classification academic, work, and social contexts.
of stress levels.  A Multi-Layer Perceptron (MLP) classifier is
 This contributes significantly to then employed to model the complex
understanding the mental health relationships between textual features and stress
consequences during internet shutdowns levels. The MLP network consists of multiple
and aids in identifying individuals who hidden layers with nonlinear activation functions,
may need targeted psychological support. enabling it to learn intricate patterns in participant
responses. The model is trained on a labeled
dataset, where stress levels are categorized from
minimal to extreme, and optimized using
backpropagation and gradient descent. Once
trained, the system can accurately classify new
responses, predicting stress intensity in real-time.
This approach provides a scalable, automated tool
to identify individuals at risk, supporting timely
mental health interventions and aligning with
global objectives for well-being.
EXISTING ALGORITHM PROPOSED ALGORITHM: -
 Support Vector Machine (SVM)  MLP Classifier with NLP
ALGORITHM DEFINITION: ALGORITHM DEFINITION:
Support Vector Machine (SVM) is a The MLP (Multi-Layer Perceptron) classifier is a
supervised machine learning algorithm used feedforward neural network used for supervised
primarily for classification and regression learning, capable of modeling complex, nonlinear
tasks. It works by finding the optimal relationships between input features and target labels.
hyperplane that best separates data points of In this system, textual data from surveys is first
different classes in a high-dimensional space, processed using NLP techniques such as
maximizing the margin between them. This tokenization, stop-word removal, and vectorization to
approach helps SVM achieve high accuracy convert text into numerical feature vectors. These
and robustness, especially in complex and non- vectors serve as inputs to the MLP, which consists of
linear problems. an input layer, one or more hidden layers with
nonlinear activation functions, and an output layer
corresponding to stress categories. During training,
the MLP uses backpropagation and gradient descent
to minimize prediction error, adjusting the weights
across layers to learn patterns in language that
indicate varying stress levels. Once trained, the
classifier can predict stress intensity from new textual
responses, enabling automated and accurate mental
health assessment.
DRAWBACKS: - ADVANTAGES: -
 High Computational Cost for Large  Ability to Learn Complex Nonlinear
Datasets Relationships
 Limited Performance with Noisy Data  Efficient Handling of Large-Scale Text Data
 Difficulty in Selecting Optimal Kernel  Flexibility with Multiple Hidden Layers
 Poor Scalability for Real-Time  Strong Generalization on Unseen Data
Applications  Scalable for Real-Time Stress Prediction
 Less Effective with Complex Nonlinear
Patterns
SYSTEM ARCHITECTURE:

Pre-
processing MLP
Collecting and Feature Predict the stress Performance
Classifier
Stress level Selection Analysis
dataset

Fig: - proposed model

 PROCESSOR : Pentium i3 Processor

SOFTWARE REQUIREMENTS

 BACK END : PYTHON