Scribd
Upload
6 views13 pages

A Machine Learning-Based Solution For Enhanced Online Security

Uploaded by

kulsum8792
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
6 views13 pages

A Machine Learning-Based Solution For Enhanced Online Security

Uploaded by

kulsum8792
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

Phishing

Website
Detection Using
Machine
Learning
A MACHINE LEARNING-BASED SOLUTION FOR
ENHANCED ONLINE SECURITY

PRACHI GUPTA-
4MH22CS109
KUSHI RAO PR-
4MH22CS074
VIDYASHREE N-
4MH22CS180
Introduction

 Phishing attacks are a major cybersecurity threat


today.
 This project leverages machine learning for real-
time detection.
 Creating a system to detect phishing websites
using machine learning.
 System analyses and classifies URLs as legitimate
or phishing
Aim and Objectives
 Aim:
 Build a high-accuracy tool to classify URLs as phishing
or legitimate.

 Objectives:
 Feature extraction from URLs (length, subdomains,
characters).
 Evaluate and compare models for best performance.
 Develop a user-friendly web interface for URL analysis.
Current Scenario of
Phishing Attacks

• Phishing Overview:
• One of the most common cyber threats today.
• Targets individuals and organizations.
• Consequences:
• Financial loss, identity theft, and data breaches.
• Challenges:
• Increased online activities make traditional detection
harder.
Machine Learning
Algorithms

• Key Algorithms Tested:


• XGBoost: High accuracy (86.7%) and efficiency.
• Decision Trees & Random Forest: Basic
classifiers.
• SVM: Helps separate phishing from legitimate URLs.
• Neural Networks: Explored for deeper analysis.
• Why XGBoost?
• Best performance.
• Handles complex data efficiently.
Technologies Used

 Machine Learning Models: XGBoost, Random


Forest, Decision Tree, SVM.
 Development Framework: Flask for user interface.
 Programming Language: Python.
 Dataset: 10,000 URLs (50% phishing, 50%
legitimate).
How the System Works

• Feature Extraction:
• Analyzes URL length, special characters, and HTTPS
usage.
• Model Analysis:
• Trained model predicts if a URL is phishing or
legitimate.
• User Interaction:
• Real-time detection through a Flask-based web
interface.
Results and Performance

 XGBoost emerged as the top model with 86.7%


accuracy.
 Features used include URL length, subdomains,
HTTPS presence.
 Real-time detection capability implemented for
instant feedback.
Understanding Phishing
Attacks

• What is a phishing attack?


• Fake websites designed to steal sensitive
information.
• How do these attacks happen?
• Links via email, text, or ads.
• Fake sites mimic legitimate websites.
Detecting Phishing URLs

•Signs to Look For:


•Very long or unusual URLs.
•Excessive subdomains or dots.
•Lack of HTTPS security.
•Suspicious characters in the URL
Effects of Phishing Attacks

• Consequences:
• Financial Loss: Unauthorized transactions.
• Identity Theft: Misuse of personal data.
• Data Breaches: Loss of sensitive organizational data.
• Use of Stolen Data:
• Sold on the dark web.
• Used for fraud or blackmail.
Prevention Measures

 Verify URLs before clicking.


 Use multi-factor authentication and strong
passwords.
 Use Browser extensions and anti-phishing tools .
 Stay educated about phishing techniques.
 Report and block phishing URLs when detected.
Conclusion

 Developed a robust system for phishing detection


using XGBoost.
 Provides real-time URL analysis to identify
phishing threats.
 Future enhancements include browser integration
and continuous updates.
 A step forward in combating phishing attacks
effectively.

You might also like