Claim Missing Document
Check
Articles

A CNN–LSTM–DQN Policy with Prioritized Experience Replay for Cost-Aware Intrusion Detection on CSE-CIC-IDS2018 Rushendra; Kalamullah Ramli; Prima Dewi Purnamasari
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2629

Abstract

Purpose – This study aims to examine intrusion detection as a policy learning problem and determine how replay strategy alone controls the operational trade-off between attack recall and alert volume under extreme class imbalance. Design/methods/approach – A controlled ablation study was conducted using a fixed CNN–LSTM feature extractor, fixed feature set, and fixed reward structure combined with a Deep Q-Network (DQN) agent. Three configurations were compared: naïve DQN without replay, uniform experience replay with a target network, and Prioritized Experience Replay (PER). Experiments used the CSE-CIC-IDS2018 dataset, consisting of 10,788,508 training flows and 2,697,128 testing flows, with attack events occurring at fewer than 70 per million flows. Performance was assessed through recall and alerts per million flows (ARMF). Findings - Supervised CNN–LSTM baselines achieved recall above 95% but generated 31,000–45,000 ARMF. Naïve DQN reduced ARMF to 383 but sharply decreased recall to 42.47%. Uniform replay improved recall to 84.95% but increased ARMF to 12,728. PER achieved the most balanced operating point, reaching 91.40% recall at 1,031 ARMF, approximately 30 times fewer alerts than the supervised CNN–LSTM reference, with a 5.91-percentage-point recall cost. Research implications/limitations – The findings indicate that replay distribution is a critical operational design variable for controlling alert volume in highly imbalanced intrusion detection settings. However, the study is limited to a fixed backbone, feature set, reward shape, and dataset. Originality/value – This study demonstrates that replay strategy can substantially reshape IDS operating points independently of model architecture or feature representation.
Reinforcement Learning-Based Adaptive Threshold Optimization for Federated Sequence-to-Sequence Anomaly Detection in IoT Network Traffic Rahmad Hidayat Hadi Subroto; Kalamullah Ramli
Journal of Vocational, Informatics and Computer Education Vol 4, No 2 (2026): June 2026
Publisher : Academic Bright Collaboration

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.66053/voice.v4i2.946

Abstract

Purpose – This study evaluates reinforcement learning (RL)-based adaptive threshold optimization for binary Internet of Things (IoT) network anomaly detection in a federated sequence-to-sequence (Seq2Seq) reconstruction model. Reconstruction-based detectors require a threshold to convert anomaly scores into benign-or-attack decisions, and a fixed threshold may not provide the most suitable operating point in federated non-IID settings.Methods – An LSTM Seq2Seq autoencoder was implemented using a controlled CICIoT2023 subset with 182,829 records and 39 numerical traffic features. Three scenarios were compared: centralized Seq2Seq with static threshold, federated Seq2Seq with static threshold, and federated Seq2Seq with RL-based adaptive threshold. Federated learning used five simulated clients, Dirichlet non-IID partitioning with α = 0.5, three local epochs, ten communication rounds, and weighted FedAvg aggregation. The RL component was implemented as an offline validation-based threshold optimizer and selected a single final threshold after federated training. Findings – Compared with federated static thresholding, the RL-based adaptive threshold improved accuracy from 92.97% to 95.71%, recall from 91.60% to 98.21%, and F1-score from 94.98% to 97.08%. FNR decreased from 8.40% to 1.79%, while FPR increased from 3.40% to 10.95%.Research implications – Threshold optimization should be treated as a decision-layer component in federated reconstruction-based IDS. The proposed pipeline may also support vocational informatics and cybersecurity education as a case study on federated learning, anomaly detection, and threshold-based IDS trade-offs.Originality – This study positions RL-based threshold adaptation as a decision-layer component in federated reconstruction-based IoT anomaly detection.
IT Maturity Level Analysis Using Framework COBIT 5 Work for Management Cyber Incident: Case Study of Company Z in ICT Field Inna Madiyaningsih; Kalamullah Ramli
Gema Wiralodra Vol. 14 No. 2 (2023): gema wiralodra
Publisher : Universitas Wiralodra

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31943/gw.v14i2.476

Abstract

Based on the report security year 2022 from company Z in the ICT sector, it was found that on the report firewall security, 96% of level threat is level critical, and 4% is level high. Report email security is session domain 27.3%; Session limits 15.35 %; Forti Guard AntiSpam-IP 3.11%, and the rest is receipt verification; directory filter and access control relay denied. With the challenge of current cyber-attacks, precious data assets need analysis to measure the the level of IT maturity that can ensure stakeholders' interest and maximize benefits and opportunities through technology information. We overcome limitations by analyzing IT maturity using COBIT 5. Research focused only on the APO13 and DSS05 process domains. A study was done to identify the problems based on results observation, checking, and use of a questionnaire in a direct manner. Measurement is done through method evaluation self and interviews deep with the IT team and all power expert who has COBIT certification 5. Analysis results show that the measurement level for the APO13 domain is 3, and for DSS05 is level 2. These results still need to be below the set level 4 target management; therefore, building a framework to monitor, track, and record security data in real time is necessary. With the build framework, Work can help lower the threat level from a critical level to a high level and increase the COBIT Maturity Level to APO13 and DSS05 according to organizational targets.
Analisis dan Penilaian Manajemen Risiko Keamanan Informasi Menggunakan SNI ISO/IEC 27005:2022 pada Aplikasi One Data Instansi XYZ Yudhi Prasetya; Kalamullah Ramli
Jurnal Informatika: Jurnal Pengembangan IT Vol 11, No 2 (2026)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v11i2.10391

Abstract

The One Data application at XYZ Agency is a strategic data portal that supports bureaucratic reform and civil service governance. The high strategic value of the managed data has implications for increasing information security risks. This study aims to design information security risk management in the One Data Application at XYZ Agency by referring to SNI ISO/IEC 27005:2022, prioritizing risk treatment, and validating the design through control alignment with the IKASANDI guidelines. The research methods include inventorying application-related assets, identifying threats and vulnerabilities, formulating risk scenarios, and analyzing and evaluating risks using consequence-likelihood criteria and a 5x5 risk matrix referring to PermenPANRB Number 43 of 2021. The design results produce a risk register covering 30 assets, 21 types of threats, and 61 risk scenarios, consisting of 33 risks requiring mitigation and 28 risks in the acceptable category. High-priority risks primarily relate to service availability, dependence on supporting infrastructure, and human factors, including potential insider threats and limited personnel availability. The risk treatment recommendations are mapped against SNI ISO/IEC 27002:2022 controls and aligned with IKASANDI controls; design validation is demonstrated by an increase in the IKASANDI score from 1,65 to 2,24. The proposed risk treatment plan also targets a decrease in the average risk score from 12,07 to 6,31, placing all scenarios in the acceptance category. The study concludes that the implementation of SNI ISO/IEC 27005:2022, validated through an increase in the IKASANDI score, results in a structured risk management design, and supports continuous monitoring and improvement of the information security management system in the XYZ Agency's One Data Application.