Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal JURNAL INSTEK (Informatika Sains dan Teknologi) Jurnal FASILKOM (teknologi inFormASi dan ILmu KOMputer)
Andri Yudha Pratama
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Electrical & Electronics Engineering Engineering

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
MODEL DETEKSI DDOS BERBASIS MACHINE LEARNING YANG EFISIEN, INTERPRETABLE, DAN SIAP IMPLEMENTASI OPERASIONAL Andri Yudha Pratama; Khalifatur Rauf; Enny Itje Sela
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 11 No 1 (2026): APRIL
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/instek.v11i1.65922

Abstract

Serangan Distributed Denial of Service (DDoS) menjadi ancaman serius bagi kontinuitas bisnis digital, sehingga membutuhkan sistem deteksi yang akurat, responsif, dan dapat diinterpretasikan. Sebagian besar penelitian terdahulu berfokus pada maksimalisasi akurasi melalui model kompleks, namun kerap mengabaikan efisiensi komputasi dan actionability yang esensial bagi implementasi real-time. Penelitian ini mengevaluasi sembilan skenario deteksi pada dataset CIC-DDoS2019 melalui kombinasi metode seleksi fitur (Pearson, ANOVA, RFE) dan algoritma machine learning (Decision Tree, Random Forest, Logistic Regression). Hasilnya mengungkapkan adanya trade-off signifikan antara kompleksitas model dan latensi deteksi. Penelitian ini mengidentifikasi Skenario E4 (RFE + Decision Tree) sebagai model terbaik berdasarkan trade-off akurasi, latensi, dan memori, dengan recall serangan 0,9999, latensi 900 µs (sekitar 38 kali lebih cepat dari Random Forest), dan efisiensi memori 5.760 Byte. Kontribusi utama penelitian ini mencakup evaluasi multi-objektif yang mengintegrasikan akurasi, latensi, memori, interpretabilitas, dan robustness; pemetaan fitur SHAP ke dalam matriks mitigasi Defense-in-Depth; serta bukti empiris trade-off antara efisiensi operasional dan ketahanan model terhadap serangan adaptif. Analisis SHAP menunjukkan keputusan model didasarkan pada fitur identitas, anomali TCP flag, dan pola idle time. Namun, uji robustness mengindikasikan kerentanan terhadap manipulasi input, menegaskan perlunya strategi mitigasi tambahan dalam kerangka Defense-in-Depth agar model tidak hanya unggul secara statistik, tetapi juga operasional dan adaptif terhadap ancaman cerdas.
H-ASICS: Desain Intrusion Detection System Adaptif Berbasis Hybrid Deep Learning untuk Infrastruktur Kritis Andri Yudha Pratama; Ujianto, Erik IH; Rianto, Rianto
JURNAL FASILKOM Vol. 16 No. 1 (2026): Jurnal FASILKOM (teknologi inFormASi dan ILmu KOMputer)
Publisher : Unversitas Muhammadiyah Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37859/jf.v16i1.11006

Abstract

The digital transformation of critical infrastructure, particularly Smart Grid and SCADA systems, has exposed new vulnerabilities to complex cyber-attacks such as False Data Injection (FDI), necessitating proactive defense mechanisms that transcend conventional approaches. Through a Systematic Literature Review (SLR) of 51 state-of-the-art studies (2022–2026), this research confirms a paradigm shift from static Deep Learning models toward adaptive, transparent, and decentralized detection ecosystems. Addressing the critical trade-off between high accuracy and operational latency, this study proposes the conceptual framework of H-ASICS (Hybrid Adaptive System for Infrastructure Critical Security). Based on a closed-loop MAPE-K architecture, H-ASICS dynamically selects the most optimal detection algorithms switching between Hybrid CNN-LSTM for complex spatial-temporal patterns and LightGBM for edge computing efficiency. Addressing the critical trade-off between high accuracy and operational latency, this study proposes the conceptual framework of H-ASICS (Hybrid Adaptive System for Infrastructure Critical Security). Based on a closed-loop MAPE-K architecture, H-ASICS dynamically selects the most optimal detection algorithms switching between Hybrid CNN-LSTM for complex spatial-temporal patterns (yielding up to 99.81% detection accuracy) and LightGBM for edge computing efficiency (reducing operational latency to under 10 ms). The superiority of H-ASICS is further reinforced by the integration of Explainable AI (XAI) and blockchain technology to guarantee the transparency of mitigation decisions and the immutability of cyber forensic data. This proposed architecture provides a strategic roadmap for next-generation security systems that are not only accurate and resilient but also highly accountable.
Co-Authors Enny Itje Sela Khalifatur Rauf Rianto Rianto Ujianto, Erik IH