WELCOME
VIDYAVARDHAKA SANGHA®, MYSURU
SRI K PUTTASWAMY FIRST GRADE COLLEGE
VVCE Campus, Gokulam 3rd Stage, Mysuru – 570002
DATA LEAKAGE TRACKING SYSTEM USING MACHINE LEARNING
UNDER THE GUIDENCE OF
Aishwarya N
Asst. Professor
Department of Computer Application
SUBMITTED BY
Rishi N (U01CM23S0016 )
Punith BJ (U01CM23S0015)
Samrudh KM (U01CM23S0014 )
Purushotham KJ (U01CM23S0066)
INDEX
• ABSTRACT
• SYSTEM ANALYSIS
• FEASIBILITY STUDY
• EXISTING SYSTEM
• PROPOSED SYSTEM
• S/W AND H/W REQUIREMENTS
• SYSTEM DESIGN
• ER DIAGRAMS
• DFD
• FLOW CHART
• FUTURE ENHANCEMENTS
ABSTRACT
• In the digital era, organizations store vast amounts of sensitive data such as personal information, financial
records, intellectual property, and confidential business documents. Data leakage—either intentional or
accidental—poses serious security threats, leading to financial loss, legal penalties, and reputational damage.
• Traditional data protection mechanisms such as access control and encryption are not sufficient to detect
insider threats or sophisticated cyberattacks. The Data Leakage Tracking System Using Machine Learning
aims to address these challenges by continuously monitoring data access patterns and identifying abnormal
behavior that indicates potential data leakage.
• The system applies machine learning techniques to analyze user activities, data usage patterns, file access
logs, and network behavior to detect suspicious activities in real time. It not only detects data leakage but also
helps track the source of the leak, enabling timely intervention and investigation.
SYSTEM ANALYSIS
• The Issue: Unstructured and structured data leaving the organization's
network via emails, USBs, cloud uploads, or printed documents.
• The Challenge: Traditional DLP (Data Loss Prevention) rules generate too
many false positives and miss sophisticated attacks.
• The Solution: An ML-driven system that learns normal behavior and
identifies anomalies to predict and prevent leaks in real-time.
EXISTING SYSTEM
• Static access control mechanisms.
• Rule-based data monitoring.
• Limited detection of insider threats.
• No behavior-based anomaly detection.
• Manual investigation of data breaches.
• High false positives or undetected leaks
PROPOSED SYSTEM
• The proposed system introduces an intelligent and automated data leakage
tracking approach:
• Continuous monitoring of data access activities.
• Machine learning–based anomaly detection.
• User behavior profiling.
• Real-time alerts for suspicious activities.
• Tracking the source of data leakage.
• Role-based access for admin and users.
• This system enhances data security and provides proactive data leakage
prevention.
HARDWARE REQUIREMENTS
• Processor : Intel i3 or higher
• RAM : 4 GB minimum
• Hard Disk : 200 GB
• Network : Secure Internet/Intranet
SOFTWARE REQUIREMENTS
• Operating System : Windows 10 / 11
• Frontend : HTML5, CSS3
• Backend : C# [Link]
• Database : MS SQL Server
• ML Framework : [Link] / Python
• IDE : Visual Studio
SYSTEM DESIGN
• Modules
• 1. User Module
• Secure user registration and authentication.
• Access authorized data resources.
• Perform data upload/download operations.
• Activity logging and monitoring.
• 2. Data Monitoring Module
• Track file access, transfers, and modifications.
• Monitor data movement across systems.
• Log user actions with timestamps.
• 3. Machine Learning Module
• Learn normal user behavior patterns.
• Detect anomalies in data access.
• Classify activities as normal or suspicious.
• Continuously improve detection accuracy.
• 4. Data Leakage Tracking Module
• Identify leaked data instances.
• Trace leak origin using logs and metadata.
• Analyze data flow paths.
• 5. Alert & Notification Module
• Real-time alerts for suspicious activities.
• Email/SMS notifications to administrators.
• Dashboard alerts and logs.
• 6. Admin Module
• Monitor system activities and alerts.
• View data leakage reports.
• Manage users and access permissions.
• Configure security policies and thresholds
SYSTEM DESIGN
• ER DIAGRAM
DFD DIAGRAM
FLOW CHART
FUTURE ENHANCEMENTS
•Integration of Deep Learning Models – Use advanced NLP and transformer
models for better context-aware sensitive data detection.
•Federated Learning Implementation – Enable secure model training across
multiple systems without sharing raw data.
•Real-Time AI-Based Risk Prediction – Implement predictive analytics to
identify potential data leaks before they occur.
•Zero-Trust Security Integration – Combine with zero-trust architecture for
continuous verification of users and devices.
•Behavioral Biometrics & Advanced UBA – Enhance insider threat detection
using keystroke dynamics and detailed user behavior analytics.
CONCLUSION
• The Data Leakage Tracking System Using Machine Learning
provides a proactive and intelligent approach to securing sensitive
data. By leveraging machine learning and behavioral analysis, the
system can detect and trace data leakage incidents in real time,
significantly reducing security risks. This solution strengthens
organizational data protection and enhances trust in digital
systems.
THANK YOU