AI-Driven Intelligent Compliance and Resilience Framework (AI-ICRF) for Critical U.S. Aviation Spare Parts and Semiconductor Supply Chains

Authors

  • Kamana Parvej Mishu College of Graduate and Professional Studies, Trine University, Angola
  • Md Shamannoor Mohiyan College of Graduate and Professional Studies, Trine University, Angola
  • Md Majadul Islam Jim Department of Information Systems, Lamar University, Texas, USA
  • Mohammad Atif Department of Industrial & Manufacturing Engineering, FAMU-FSU College of Engineering 2005 Levy Ave
  • Samia Ara Chowdhury Department of Information Technology, St Francis College, Brooklyn, New York, USA

Keywords:

Aviation regulatory compliance; Semiconductor supply chain resilience; machine learning; NLP; Compliance-Resilience Index; ensemble methods; explainable AI.

Abstract

In the U.S., the FAA oversees aviation regulatory compliance, and the International Civil Aviation Organization (ICAO) supports aviation sector partners globally to manage operational risks, including semiconductor supply chain disruptions. However, traditional governance models, which rely on time-consuming manual audits periodically, are completely inadequate in the fast-paced, interdependent supply chains of today. Abstract: The AI-Driven Intelligent Compliance and Resilience Framework (AI-ICRF) develops a four-layer intelligent system using BERT-based NLP, ensemble machine learning ( Random Forest, XGBoost ), and LSTM deep learning. The framework is trained on 856,292 records from all three federal datasets (2012 to 2022) and is internally validated using rigorous statistical methods with a 97.1% compliance prediction accuracy. Statistical validation: Ten-fold cross-validation, paired t-tests, one-way ANOVA, McNemar's test, DeLong AUC comparison, bootstrap confidence intervals (n = 10,000 iterations), and SHAP-based explainability. Combing all three modalities, AI-ICRF Ensemble predicts compliance with 97.1% accuracy (AUC-ROC = 0.984, 95% CI: 0.981–0.987), 22.9 percentage points better than rule-based baselines (p 24.6, p < 0.001 for all comparisons). HiBIS is a quantitative prioritization process based on the composite Compliance-Resilience Index (CRI) which shows very large discriminative validity against 42 months of historical FAA enforcement data (Cohen's d = 4.63, 95% CI: 3.71–5.55). SHAP analysis indicates a lack of NLP-derived documentation as the third most important predictor (mean |SHAP| = 0.187). This research establishes a scalable and empirically validated computational intelligence architecture that explicitly governs proactive regulatory compliance in safety-critical aerospace industrial ecosystems and has direct policy relevance to the CHIPS and Science Act framework and the FAA Next Generation Safety Oversight initiative.

Published

27-12-2022

How to Cite

Kamana Parvej Mishu, Md Shamannoor Mohiyan, Md Majadul Islam Jim, Mohammad Atif, & Samia Ara Chowdhury. (2022). AI-Driven Intelligent Compliance and Resilience Framework (AI-ICRF) for Critical U.S. Aviation Spare Parts and Semiconductor Supply Chains. Well Testing Journal, 31(2), 240–271. Retrieved from https://welltestingjournal.com/index.php/WT/article/view/313

Issue

Section

Original Research Articles

Most read articles by the same author(s)

Similar Articles

<< < 8 9 10 11 12 13 14 15 16 > >> 

You may also start an advanced similarity search for this article.