Simulation Code: Boolean Differential Analysis with Hybrid Density-Partitioning Clustering for Enhanced Security Risk Assessment
Year: 2026
Type: software
Abstract: This repository contains the complete Python simulation package accompanying the paper "Boolean Differential Analysis with Hybrid Density-Partitioning Clustering for Enhanced Security Risk Assessment" by Gunel Babanli and Rana Amirova (Azerbaijan Technical University). The package reproduces all figures, tables, and numerical results presented in the paper: - Boolean differential operators (∂f/∂xᵢ, max∨, min∧) for f(x₁,x₂,x₃) = (x₁·x₂) ⊕ x₃- Linear and sigmoid AI risk models (Figures 2–3)- Critical variable x₃ analysis and 3D risk surfaces (Figures 4–5)- Preprocessing pipeline: binarization → Boolean feature augmentation → normalization- K-Means clustering (k=3) with quantization mapping (Table 2)- DBSCAN density-based clustering (ε=0.30, minPts=2)- Hybrid K-Means + DBSCAN method with Silhouette Score evaluation (Table 3)- Neuro-symbolic training comparison: Standard AI vs. Boolean-guided model (Table 1) Requirements: Python 3.8+, numpy, matplotlib, scikit-learn, scipy Run: python run_all.py
Authors: Amirova, Rana, Babanli, Gunel
Institution: Azerbaijan Technical University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7166345537