XAI and Statistical Analysis for Reliable Intrusion Detection in the UAVIDS-2025 Dataset: From Tree to Hybrid and Tabular DNN Ensembles

2026-05-13

AI Summary

This study addresses the challenges of opaque decision-making in complex models for UAV intrusion detection and the misclassification of Wormhole and Blackhole attacks caused by overlapping density supports. To overcome these issues, the authors propose a high-accuracy detection framework that integrates XGBoost, deep neural networks, and a hybrid stacking ensemble, validated on the UAVIDS-2025 dataset. Innovatively combining SHAP interpretability analysis, Westfall–Young multiple testing correction, KDE bandwidth optimization, and Jensen–Shannon divergence, the work systematically uncovers the masking characteristics of these attacks and the root causes of misclassification. Through stratified 10-fold cross-validation and visual statistical analysis, the approach not only significantly enhances detection reliability but also, for the first time, elucidates the伪装 mechanisms and decision boundaries distinguishing Wormhole and Blackhole attacks at a mechanistic level.

Abstract

During the last few years, the term Mechanistic Interpretability, a specific area, under the umbrella of explainable artificial intelligence (XAI), has been introduced, to explain the decisions made by complex machine learning (ML) models in critical systems like UAV intrusion detection systems (UAVIDS). In this paper, we apply best-practices for data pre-processing and examine a wide range of tree-ensembles, deep neural networks, hybrid stacking models and the latest ensemble neural networks to detect intrusions in UAV, with stratified 10-fold cross validation. With our top-performing model, XGBoost, we proceed to Shapley Additive explanations (SHAP), to analyze the global and local feature importances and understand which features, each attack targets, to mimic normal traffic and where the misclassifications occur. Furthermore a distribution analysis follows, by visually comparing violin plots and the curves of kernel density estimations. With the Westfall-Young permutation test for multiple comparisons, the Bandwidth optimization of the KDEs and the selection of Jensen-Shannon Distance for the test, we discover the true causes of false predictions, observed in Wormhole and Blackhole attacks in UAVIDS-2025. The findings provide robust, reliable and explainable models for UAV intrusion detection, along with statistical insights, which capture and clarify the masked nature of the attacks, regarding the challenge of Density Support Intersection, between these attacks, in this dataset.

Problem

• UAV intrusion detection
• explainable AI
• false predictions
• Density Support Intersection
• attack mimicry

Innovation

• Explainable AI (XAI)
• Mechanistic Interpretability
• Statistical Hypothesis Testing
• Ensemble Neural Networks
• UAV Intrusion Detection