Impact Factor
Call For Paper
Volume 12 Issue 07
July 2026
Download Paper Format
Copyright Form
Author(s)
Abstract
Data Leakage Is One Of The Most Critical Challenges In Machine Learning Systems, Leading To Unrealistic Model Performance And Poor Generalization In Real-world Applications. Leakage Occurs When Information From Outside The Training Dataset Is Inadvertently Used During The Model Training Process, Causing Biased Predictions And Overly Optimistic Evaluation Metrics. Detecting Such Leakage Before Model Development Is Essential For Building Reliable And Robust Machine Learning Systems. This Paper Proposes A Data Leakage Detection And Intelligent Data Preprocessing System That Automatically Identifies Potential Leakage Sources In Datasets Prior To Model Training. The System Integrates Dataset Profiling, Leakage Detection, Preprocessing Techniques, And Visualization Tools Within A Flask-based Web Application. Users Can Upload Datasets, Analyze Data Quality, Detect Different Types Of Leakage Such As Target Leakage And Temporal Leakage, And Apply Safe Preprocessing Operations. The System Also Provides Interactive Data Visualizations And Exports A Cleaned Dataset Ready For Machine Learning Tasks. By Combining Leakage Detection With Automated Preprocessing, The Proposed Solution Improves Model Reliability, Reduces Human Error, And Enhances The Overall Machine Learning Workflow.
Keywords
Paper ID
IJSARTV12I5105356
Publication Date
May 14, 2026
Research Area
Artificial Intelligence And Data Science