Software Requirements Specification for

Fake News Prediction Using Machine Learning

Submitted by
Group 7
AP21110011580
AP21110011581
AP21110011582
Table of Contents
Table of Contents.......................................................................................................................................1

1. Introduction.........................................................................................................................................2

1.1 Purpose.................................................................................................................. .........................2

1.2 Document Conventions..................................................................................................... ..............2

1.3 Intended Audience and Reading Suggestions..................................................................................2

1.4 Product Scope............................................................................................................ .....................2

1.5 References............................................................................................................... .......................2

2. Overall Description....................................................................................................... ......................2

2.1 Product Perspective...................................................................................................... ...................2

2.2 Product Functions........................................................................................................ ...................3

2.3 User Classes and Characteristics.....................................................................................................3

2.4 Operating Environment.................................................................................................... ...............3

2.5 Design and Implementation Constraints.................................................................................... .....3

2.6 User Documentation.......................................................................................................................3

2.7 Assumptions and Dependencies............................................................................................. .........3

3. External Interface Requirements........................................................................................... ............4

3.1 User Interfaces................................................................................................................................4

3.2 Hardware Interfaces...................................................................................................... ..................4

3.3 Software Interfaces...................................................................................................... ...................4

4. System Features........................................................................................................... ........................4

4.1 Use Cases........................................................................................................................................5

4.2 User Authentication and Ingestion..................................................................................................5

4.3 Preprocessing and Feature Extraction.............................................................................................5

4.4 User Interface............................................................................................................................. ......5

5. Other Nonfunctional Requirements...................................................................................................5

5.1 Performance Requirements............................................................................................................. 5

5.2 Safety Requirements...................................................................................................... .................5

5.3 Security Requirements....................................................................................................................5

5.4 Software Quality Attributes............................................................................................................6

5.5 Project and user documentation......................................................................................................6

6. Other Requirements................................................................................................................. ...........6

Appendix A: Glossary......................................................................................................... .......................6

Appendix B: Analysis Models...................................................................................................................7

1 Introduction
1.1 Purpose
This Software Requirements Specification (SRS) document outlines the functional
and non-functional requirements for developing a Fake News Prediction system using
machine learning. This system aims to detect and classify news articles as either
authentic or fake, contributing to mitigating misinformation and disinformation.

1.2 Document Conventions

Terminology:
Real News: Genuine and reliable news articles.
Fake News: Misleading or false information presented as news.
Abbreviations:
ML: Machine Learning
SRS: Software Requirements Specification

1.3 Intended Audience

This document is intended for the development team, project managers, quality
assurance team, and stakeholders involved in the creation of the Fake News Prediction
system. It provides a comprehensive understanding of the requirements and guidelines
for the successful implementation of the project.

1.4 Product Scope

The project's scope involves creating an effective Fake News Detection System
utilizing machine learning models, including Decision Tree, Random Forest, SVM,
Naive Bayes, Logistic Regression, and passive-aggressive classifier. The system will
classify news articles as genuine or fake, using datasets sourced from Kaggle,
specifically the "Fake and Real News" and "Fake News Classification" datasets, for
model training and evaluation.
The objective is to contribute to the fight against misinformation by providing a
reliable and adaptable tool for identifying fake news articles and leveraging advanced
machine learning and NLP techniques.

2. Overall Description
2.1 Product perspective
The perspective of the Fake News Detection project lies in deploying machine learning
algorithms that involve training and integration of natural language processing (NLP)
algorithms play a vital role in extracting essential features from textual content,
enhancing the system's ability to identify patterns associated with misinformation.

2.2 Product functions

 The product functions of the fake news prediction are as follows:
1. Text preprocessing for feature extraction
2. Machine learning model training
3. Real-time news classification
4. User feedback and model retraining

2.3 User classes and characteristics

Readers: Users who interact with the system to check the authenticity of
news articles.
Administrators: Users responsible for system maintenance and monitoring.

2.4 Operating environment

The system will run on cloud infrastructure, utilizing services like AWS or Google
Cloud. It requires Python and relevant ML libraries. Users will access the system
through a web interface or API. The system must support various browsers.

2.5 Constraints
The implementation of the Fake News Detection project using ML models
encounters certain constraints that influence the development and deployment
processes. These constraints include:
Computational Resources: Limited computational resources may impact the
scalability and efficiency of training complex ML models, especially when dealing
with large datasets. High computational requirements can pose challenges for real-
time processing and model optimization.

2.6 Documentation
The documentation includes detailed processes for data collection, preprocessing,
model training, and evaluation, with an emphasis on confidence score calculation. A
user-friendly interface is developed, and security measures are implemented to
ensure data integrity. The documentation provides stakeholders with a
comprehensive understanding of the Fake News Detection system, covering data
aspects, algorithmic intricacies, and user interaction functionalities.

2.7 Assumptions and Dependencies

The assumption is made that the Kaggle datasets, "Fake and Real News" and "Fake
News Classification," maintain integrity and provide an unbiased representation of
news articles, forming a solid foundation for model training. Another critical
assumption involves compliance with ethical considerations and data privacy
regulations during the collection and utilization of news articles, ensuring user
information is safeguarded. A key dependency lies in the performance of the NLP
algorithms for feature extraction, assuming their ability to capture nuanced linguistic
patterns effectively
3. External Interface Requirements
3.1 User interfaces
The interface displays the confidence scores generated by the models, offering a clear
indication of the likelihood of the news being fake. Elements such as charts or graphs
enhance result interpretation. User-friendly features ensure accessibility and ease of
use, promoting a positive experience. The interface design is adaptable and
compatible across various devices and browsers, ensuring widespread accessibility for
users.

3.2 Hardware interfaces

The hardware interface supports data storage for the Kaggle datasets, "Fake and Real
News" and "Fake News Classification," and ensures seamless integration with natural
language processing algorithms for feature extraction. As a software-driven project,
the hardware interface focuses on providing a robust computational environment,
facilitating the effective implementation and performance of the Fake News Detection
system.

3.3 Software Interfaces

The software interface for the Fake News Detection project is implemented using
programming languages such as Python and ml libraries:
▪ Numpy
▪ Pandas
▪ Scikit-learn
▪ Matplotlib
▪ seaborn
▪ Tensor flow
The software interface ensures seamless integration with Kaggle datasets, "Fake and
Real News" and "Fake News Classification," facilitating data preprocessing and
model training. Additionally, the interface accommodates the incorporation of NLP
algorithms for feature extraction, possibly utilizing libraries like NLTK.

4. System Features
The Fake News Prediction system features secure:
▪ user authentication
▪ real-time news article ingestion
▪ advanced text preprocessing for enhanced data quality.
It incorporates a robust machine learning model with confidence scores for precise
classification, complemented by detailed reporting and comprehensive logging. The
intuitive user interface showcases prediction results, while administrative
 functionalities enable efficient system management. Together, these features form a
powerful and streamlined solution for identifying and addressing misinformation.

4.1 User Authentication and Ingestion

Enable secure user authentication with options for password recovery and multi-factor
authentication. Ingest news articles from diverse sources, ensuring the system's ability
to handle real-time data.

4.2 Preprocessing and Feature Extraction

Implement text preprocessing techniques, including tokenization, stemming, and stop-
word removal, to enhance data quality. Extract relevant features from news articles to
provide valuable input to the machine learning model.

4.3 Machine Learning Model

Develop a robust machine learning model capable of accurately classifying news
articles as authentic or fake. Confidence Score Provide a confidence score alongside
predictions to indicate the model's certainty level. Reporting and Logging Generate
detailed reports on system performance, including accuracy metrics, and maintain
comprehensive logs for auditing and debugging.

4.4 User Interface

Design a better and user-friendly interface for users to interact with the system,
displaying prediction results and relevant information. The system includes
functionalities for system configuration, user management, and model retraining. These
features collectively empower the Fake News Prediction system to effectively analyze,
classify, and report on news articles, and identify and predict the misinformation.

5. Other Nonfunctional Requirements

5.1 Performance Requirements
▪ Classify news articles within seconds to ensure real-time responsiveness.
▪ Handle a minimum of 1000 concurrent users to accommodate varying usage
loads.
5.2 Safety Requirements
▪ Implement measures to ensure no data leakage during the training or
classification processes.
5.3 Security Requirements
▪ Implement secure user authentication protocols to safeguard user access.
▪ Encrypt sensitive data during transmission to protect against unauthorized
access.

5.4 Software Quality Attributes

 ▪ Achieve a classification accuracy rate of 95% to ensure the system's
effectiveness. Regularly update machine learning models to adapt to
evolving news patterns, maintaining relevance and accuracy.
5.5 Project and User Documentation
▪ Maintain up-to-date documentation on system architecture and APIs to
facilitate system understanding and future developments.
▪ Provide a comprehensive user manual explaining how to use the system,
ensuring users can navigate and utilize its functionalities effectively.

6. Other Requirements
▪ Adhere to ethical standards, promoting fairness in system classification.
▪ Ensure scalability for increased data and user traffic in datasets.
▪ Use modular components for a scalable architecture.
▪ Support common web browsers and major operating systems.
▪ Implement continuous monitoring for system optimization.

Appendix A: Glossary
▪ NLP: Natural Language Processing
▪ ML: Machine Learning
▪ SVM: Support Vector Machine
▪ RF: Random Forest
▪ DT: Decision Tree
▪ LR: Logistic Regression

Appendix B: Project Life Cycle

1) Limitation/Disadvantages
▪ System effectiveness is influenced by biases in training data and imbalances in
authentic and fake news.
▪ Adapting to emerging fake news types poses a continual challenge, requiring
vigilance.
▪ Challenges in classifying nuanced or context-dependent articles, necessitating
ongoing refinement.

2) Application
It integrates machine learning for the secure and efficient evaluation of news articles.
The system accepts text or URL submissions for real-time analysis, presenting prompt
predictions with confidence scores. This application alerts users to potentially
misleading content, maintaining a historical record of predictions and feedback.
Educational resources foster critical thinking. Regular model updates and automated
retraining adapt to evolving misinformation patterns, contributing to informed
decision-making in the battle against fake news.