Replication Package for Hybrid Vulnerability Detection Framework

This repository contains the complete replication package for the manuscript "An Explainable Hybrid Learning Framework for Vulnerability Detection to Enhance Software Dependability" submitted for peer review.Contents:Preprocessed BigVul dataset features (semantic and structural embeddings)Trained model checkpoints (CodeBERT, GCN, Random Forest)Source code for feature extraction, training, and evaluationExperimental results and statistical analysis scriptsConfiguration files for reproducibilityAll experiments use fixed random seed (42) for deterministic results.Note: This is an anonymous sharing link for double-blind peer review. Full author attribution will be added upon manuscript acceptance.