ResearchRadar
ResearchRadar
So funktioniert esQuellenAnwendungsfälleWöchentliches Radar
ENESCADEZHJAHIAnmeldenKostenlos starten
LIVE-VORLAGE · NUR LESENKein Konto erforderlich

Cybersicherheit

Sicherheitsforschung, Bedrohungserkennung und Evaluation in Preprints und Fachzeitschriften.

Suchanfrage“security detection”
Letzte 30 Tage4 Quellen15 Ergebnisse

Quellen in diesem Radar

Wähle eine oder mehrere Quellen, um die Ergebnisse zu filtern.

15 Ergebnisse
OA
VorabveröffentlichungOpenAlex·14.11.2026

An Efficient GRU-Based Intrusion Detection and Data Security Framework for IoT Networks Using Blockchain

Abstract The Internet of Things (IoT) has transformed numerous sectors by facilitating smart devices for collecting and exchanging data. However, the diversity and scale of IoT deployments introduce significant security risks. This paper proposes a generalizable and robust intrusion detection and data security system for IoT using KCCCR-GRU and RFIECC with blockchain. The approach integrates advanced device authentication, feature optimization, and deep learning-based attack detection, with secure data integrity ensured by blockchain. To validate generalizability, the system is evaluated on multiple benchmark IoT datasets with varying attack vectors and device types. Comparative analysis includes recent state-of-the-art IoT intrusion detection systems, such as those based on deep reinforcement learning and graph neural networks. Beyond accuracy, we assess detection latency, resource usage, and scalability to reflect real-world deployment needs. The blockchain component is described in detail, including consensus, update frequency, and system integration. We also discuss deployment considerations, such as compatibility, cost, and integration challenges. The proposed system achieves high accuracy and practical efficiency, demonstrating its suitability for diverse, large-scale IoT environments.

DOI
10.21203/rs.3.rs-11289479/v1 ↗
Autoren
Kshitiz Saxena, Dhowmya Bhatt
Zeitschrift / Publikationsort
Nicht bereitgestellt
Publikationstyp
Vorabveröffentlichung
Status
Vorabveröffentlichung
Veröffentlichungsdatum
14. November 2026
Veröffentlichte Version
In den Metadaten der Quelle nicht verknüpft
Cybersicherheit97% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·10.11.2026

Intelligent Detection of Securities Market Manipulation Using Multimodal Financial Data

Market manipulation through securities involves increasingly sophisticated mechanisms such as irregular trading, selective information disclosures, and joint activities in different securities, posing threats to surveillance relying solely on a particular information source due to both false positives and late identification of manipulations. We propose a multimodal framework for the detection of market manipulation on China's A-share market by fusing one-minute market microstructure time series, corporate announcements, financial news, and evolving relationships among securities. Enforced cases by China Securities Regulatory Commission from 2018–2025 were combined with unenforced stock-period observations and temporal, textual, and relational features are extracted using Temporal Convolutional Network, MacBERT encoder, and Graph Attention Network, respectively, followed by cross-modal attention to generate a risk score. Our framework achieved macro-F1 of 0.872 ± 0.009, AUROC of 0.946 ± 0.006, and AUPRC of 0.821 ± 0.012 on the unseen data, outperforming the TCN-only model in market only by a wide margin of 0.036 ± 0.007 and 0.028 ± 0.006 for removing text and graph information, respectively. In the 2025 market surveillance simulation, reviewing the top 10% of ranked observations led to detecting 0.643 ± 0.031 of manipulations with 0.074 ± 0.009 false-positive rate.

DOI
10.54254/2755-2721/2027.38689 ↗
Autoren
Yanying Li
Zeitschrift / Publikationsort
Applied and Computational Engineering
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
10. November 2026
Cybersicherheit97% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·10.11.2026

A hybrid framework for web applications security vulnerabilities detection

Abstract The threat landscape for modern web applications is constantly changing, and traditional security approaches struggle with comprehensive vulnerability detection. Black-box testing, although widely adopted, suffers from limited path coverage and static payload limitations. Model-Based Testing (MBT) provides systematic exploration but requires substantial effort and may miss implementation-specific vulnerabilities. This research introduces and validates a hybrid methodology that integrates black-box testing and MBT through a six-level enhancement framework incorporating authentication mechanisms, client-side crawling, custom payload optimisation, and scope refinement. To demonstrate practical implementation and validate theoretical foundations, the methodology was applied as a case study using OWASP Zed Attack Proxy (ZAP) against OWASP Juice Shop, with manual verification of 3 discovered vulnerabilities and detailed statistical analysis across multiple test iterations. The results demonstrate substantial improvements in vulnerability detection effectiveness, with strong statistical evidence. The case study implementation revealed critical tool reliability issues requiring organisational attention, including systematic documentation of interface reporting discrepancies within automated scanners affecting vulnerability count accuracy, necessitating multi-verification approaches for organisational security assessment. The research contributes (1) a systematic integration framework, designed around tool-agnostic principles and validated through a single-tool case study, enabling incremental adoption across diverse security testing platforms, (2) a comprehensive empirical validation methodology with statistical rigour, and (3) practical implementation guidance including reproducible automation scripts adaptable to various organisational contexts. The framework’s design principles and progressive enhancement approach provide a foundation for broader adoption beyond the specific ZAP implementation demonstrated in this validation study. The sixth level achieved a detection rate of 23.7 critical vulnerabilities compared to 2 critical vulnerabilities in baseline testing, representing a 1085% increase (note: based on $$n=3$$ n = 3 iterations per configuration; reported as exploratory rather than confirmatory). Total vulnerability detection increased from 583 to 4427, a 659% improvement. This demonstrates the cumulative effectiveness of systematic enhancement approaches in automated security testing within the single-application case study evaluated here.

DOI
10.1007/s10207-026-01346-7 ↗
Autoren
James A. Stewart, Gheida Shahrour, Amro Al-Said Ahmad
Zeitschrift / Publikationsort
International Journal of Information Security
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
10. November 2026
Cybersicherheit97% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·06.11.2026

AN EXPLAINABLE AI AGENT BASED ASSISTANT FOR REAL-TIME CYBER THREAT DETECTION AND PERSONALIZED SECURITY RECOMMENDATIONS

In the era of escalating cyber threats, traditional black-box intrusion detection systems (IDS) suffer from lack of transparency, high false-positive rates, and inability to provide actionable guidance to diverse users.This paper proposes an Explainable AI (XAI) AgentBased Assistant that unifies real-time threat detection with personalized security recommendations.The system integrates a hybrid CNN-LSTM detection backbone optimized for spatialtemporal network traffic analysis, achieving 98.7% accuracy and 98.6% F1-score on benchmark datasets (CICIDS2017, NSL-KDD) with sub-50ms explanation latency through optimized SHAP and LIME implementations using INT8 quantization and LRU caching.A Dynamic Explainable Framework (DEF) couplesdetection with explanation-aware alert prioritization, reducing residual false positives by 50% while improving macro precision by 2.8 points.The Personalization Enginemodels user risk appetite, historical behaviour, and role context to generate adaptive mitigation playbooks.A humanin-the-loop feedback loop enables continuous learning and trust calibration.Experimental evaluation demonstrates sustained throughput of 315 alerts per second and analyst trust improvement of 31% when explanations are presented.The architecture is designed for SOC deployment, compliance auditing, and edge applicability.

DOI
10.58257/ijprems55031 ↗
Autoren
Nicht bereitgestellt
Zeitschrift / Publikationsort
International Journal of Progressive Research in Engineering Management and Science
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
6. November 2026
Cybersicherheit97% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·13.11.2026

Basis Desk EVM & Multi-Chain Smart-Contract Security, Honeypot Detection & Renounced-Ownership Calibration Dataset (2026 ? Volume 2)

Basis Desk Quantitative Crypto Market Intelligence, Spot ETF Telemetry, EVM Contract Security & Exchange-Rate Benchmarks (2026) (https://basisdesk.news/ & https://basisdesk.news/research) ? Principal Investigator: Julian Mercer (ORCID: 0009-0003-1162-170X). Crypto-to-Fiat & Satoshi Converter Hub: https://basisdesk.news/convert ? 1 BTC to USD Live Converter (/convert/1-btc-to-usd) ? 1,000 Sats to USD Calculator (/convert/sats-to-usd, /convert/10000-sats-to-usd) ? ETH to EUR Converter ? SOL to USD ? USDT to USD ? XRP to USD Smart Contract Security Scanner & Token Audits: https://basisdesk.news/tokens ? PEPE Ethereum Contract Audit ? USDT Ethereum Contract Audit ? USDC Base Chain Contract Audit Spot Crypto ETF & Institutional Yield Tracker: https://basisdesk.news/etfs ? BlackRock IBIT ETF ? Fidelity FBTC ETF ? YieldMax MSTY Option Income ETF ? NEOS BTCI Bitcoin High Income ETF ? Strategy STRC Preferred ? Strategy STRF Preferred Historical Daily Market Close Archive & Whale Flow: https://basisdesk.news/crypto-market-today/archive ? On-Chain Whale Tracker ? Academic Research & Working Papers

DOI
10.7910/dvn/stqmmb ↗
Autoren
Julian Mercer
Zeitschrift / Publikationsort
Harvard Dataverse
Publikationstyp
dataset
Status
Veröffentlicht
Veröffentlichungsdatum
13. November 2026
Cybersicherheit93% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·09.11.2026

Semi-Automated Detection of Gaps in LLM Security Knowledge

Large language models (LLMs) are increasingly used for a range of software, hardware and human-centered security tasks. Consequently, LLM performance on security tasks is an active area of measurement and research, often with a focus on identifying areas in which LLM security “knowledge” may be insufficient. Popular strategies for identifying LLM security knowledge gaps include building corpora of challenge questions or task benchmarks, strategies that require substantial manual work and security expertise to design and execute. We introduce a partially-automated method for assessing LLM knowledge of a security area. The method uses authoritative information from Consumer Protection Agencies (CPAs) to identify instability in LLM responses that can be indicative of knowledge gaps. We demonstrate the method for 2 security topics, identity theft and impostor scams, and 5 LLMs in 2 leading LLM families, Gemini and GPT, using publicly available information about identity theft and impostor scams from 6 CPAs. The method distinguishes between models that have and don’t have sufficient knowledge to accurately identify the security topics in text narratives. Two models, GPT 5.5 and Gemini 3.1 Pro, demonstrate solid knowledge of impostor scams, whereas all models appear to have gaps in knowledge of identity theft.

DOI
10.1145/3847352.3848102 ↗
Autoren
Shufan Chai, Liangliang Sun, Jessica Staddon
Zeitschrift / Publikationsort
ACM Workshop on Artificial Intelligence and Security (AISec)
Publikationstyp
Konferenzbeitrag
Status
Veröffentlicht
Veröffentlichungsdatum
9. November 2026
Cybersicherheit93% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·09.11.2026

Harnessing AI and Machine Learning for Enhanced Cloud Security: Implications for Data Integrity, Threat Detection, and Risk Management

The quick assembly of dispersed figuring has in a general sense changed the mechanized scene, fetching relationship to store and regulate data with added created efficiency and flexibility. In any case, it has also expressed the new security challenges related to the data reality, crisis revelation and the dancing to the heads. This paper researches the ability of man-made thinking (man-made knowledge) and computer-based intelligence (ML) to further develop Cloud security, data, ensuring decency, identifying complex risks and focuses on its work in further directing stakes. By using man-made knowledge and ML, affiliation can encourage active safety procedures that automatically give the risk to the risk, supporting the overall security status of cloud conditions, automatic evidence, irregularity fields and reaction devices. This study highlights the major types of progress in man-made consciousness-driven cloud safety propels and examines its ideas to protect the data from honesty, identify risks, and lighten the related bets with dialogue.

DOI
10.1201/9781042029907-33 ↗
Autoren
Charu Singh, Deepti Sharma
Zeitschrift / Publikationsort
CRC Press eBooks
Publikationstyp
book chapter
Status
Veröffentlicht
Veröffentlichungsdatum
9. November 2026
Cybersicherheit93% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·08.11.2026

A fragile reversible watermarking technique for medical imaging security using prediction error expansion and Canny edge detection

This paper develops a fragile, reversible watermarking technique to detect accidental and intentional alterations in DICOM medical images. A fragile method is adopted to detect any manipulation applied to medical images, while reversibility ensures the recovery of the original unaltered image at extraction. The proposed approach automatically segments medical images into Region of Interest (ROI) and Region of Non-Interest (RONI), extracts authentication data from the DICOM header, and generates a digital signature of the image. The method combines integrity and authentication data and inserts them into the edge areas within the ROI, implementing the Prediction Error Expansion (PEE) technique and Canny edge detection. During extraction, the embedded watermark and the original unmanipulated image are restored without knowing the pixel locations used to hide the watermark data. The method was assessed using twelve 16-bit DICOM images, including MRI and CT, with a 512x512 pixel size. The experimental results yielded an average PSNR of 97.19 dB and an average SSIM of 0.999996, with an average watermark length of 7835 bits per image. The proposed method also fully met the reversibility requirement by recovering both the original unaltered images and the embedded watermark data. Fragility evaluation showed that manipulations of watermarked images can be detected via alterations in the retrieved watermark. The achieved results demonstrate the effectiveness of the proposed technique in maintaining visual image quality while ensuring reversibility, integrity, and authenticity in the conducted experiments.

DOI
10.55670/fpll.futech.5.4.14 ↗
Autoren
Asaad F. Qasim, Ahmed Hussain Ali
Zeitschrift / Publikationsort
Future Technology
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
8. November 2026
Cybersicherheit93% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·14.11.2026

AI-Based Intrusion Detection and Automated Response System for Enterprise Networks

Enterprise networks are increasingly exposed to cyber threats such as distributed denial-of-service attacks, brute-force login attempts, port scanning, SQL injection, unauthorized access, and abnormal service requests.Traditional intrusion detection systems are useful for monitoring network activity, but many of them mainly generate alerts and depend on manual security intervention for mitigation.This manual delay can increase the impact of attacks, especially in environments where business services, authentication servers, databases, and web applications must remain continuously available.This paper proposes an AI-based intrusion detection and automated re-sponse system for enterprise networks.The proposed system uses simulated enterprise network traffic to represent normal user activity and malicious behavior.Important features such as source IP frequency, destination port usage, request count, failed login attempts, payload pattern, protocol type, access status, and request rate are extracted from traffic logs.A machine learning model is trained to classify traffic into normal and malicious categories such as DDoS, brute force, port scan, SQL injection, and unauthorized access.After detecting the attack type, an automated response engine applies suitable mitigation actions such as IP blocking, request rate limiting, account locking, malicious request rejection, and alert generation.The system also includes an administrator dashboard for viewing logs, attack severity, response status, blocked IP addresses, and visual summaries.The proposed framework demonstrates how AI-based detection and automated response can reduce manual effort and improve enterprise network resilience.

DOI
10.56726/irjmets104676 ↗
Autoren
Nicht bereitgestellt
Zeitschrift / Publikationsort
International Research Journal of Modernization in Engineering Technology and Science
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
14. November 2026
Cybersicherheit92% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·13.11.2026

Artificial Intelligence for Intrusion Detection and Prevention: A Systematic Review of Methods, Challenges, and Future Directions

This systematic review synthesizes AI-based techniques for detecting and preventing cyberattacks published between 2019 and 2025. Following PRISMA 2020, we searched IEEE Xplore, Scopus, and the ACM Digital Library (550 records) and included 74 studies after independent dual screening; study quality was appraised with a modified JBI checklist, the risk of bias of the review was assessed with ROBIS, and the review was not registered. We include a higher-level taxonomy of present methods divided into seven categories: (1) traditional machine learning algorithms; (2) deep learning models; (3) hybrid intelligence frameworks; (4) blockchain-based systems; (5) reinforcement learning approaches; (6) ensemble learning methods; and (7) explainable AI (XAI) techniques. Reported accuracies are high in all seven categories (category medians between 95.3% and 99.9%), but individual results range from 58.17% to 100% and were obtained with different datasets, data splits, and evaluation protocols, so they do not support a ranking of the categories. Latency and resource consumption are rarely reported and, when reported, were measured under non-comparable conditions. Automated prevention actions (blocking, rate limiting, or network reconfiguration) are evaluated far less often than detection, although false positives are more costly when they trigger blocking. Targeted machine learning methods report strong performance in phishing detection (96.83–98.3% accuracy in the tabulated studies). This study also identifies several major areas needing improvement, including a predominant reliance on legacy benchmark datasets (NSL-KDD, KDD99, UNSW-NB15), which were used by 25 of the 48 studies with identifiable datasets (52.1%) and a limited capability to interpret “black box” AI models. Accordingly, we suggest that future research use more realistic test environments and recent datasets, develop lightweight and adaptive AI models, and adopt a standardized minimum reporting protocol (accuracy, precision, recall, F1-score, false positive rate, latency, resource usage, and statistical validation), for which we provide a concrete proposal and a list of recent datasets. The intention of this study is to assist cybersecurity practitioners in identifying and implementing effective AI-based security solutions in relation to present-day threats.

DOI
10.3390/network6040102 ↗
Autoren
Hanan Moufid, Mohamed El Ghazouani, Moulay Ahmed El Kiram
Zeitschrift / Publikationsort
Network
Publikationstyp
Übersichtsartikel
Status
Veröffentlicht
Veröffentlichungsdatum
13. November 2026
Cybersicherheit92% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·13.11.2026

EV Charging Station Security Ontology and Knowledge Graph: Ontology Models, SPARQL Queries, and Security Observation Dataset

This repository contains the ontology models, knowledge graph resources, SPARQL queries, and security observation dataset developed for EV Charging Station (EVCS) cybersecurity research. The collection includes: An OWL/Turtle-based ontology representing security attacks, risks, severity levels, vulnerabilities, communication protocols, system components, and mitigation mechanisms in EV charging infrastructure. An ontology incorporating cybersecurity concepts and countermeasure knowledge, including references to established cybersecurity frameworks and knowledge bases. A dataset containing 6,415 system observations represented as RDF triples, including 4,932 attack-labelled observations and 1,483 normal-labelled observations. Each observation includes four statistical features: min_top, std_top, min_stat, and min_time_diff. A collection of competency questions and SPARQL queries for statistical analysis, attack-risk-severity relationships, mitigation analysis, and knowledge graph exploration. Python scripts demonstrating ontology construction, loading, reasoning, SPARQL querying, and ontology visualisation using RDFLib and Owlready2. These resources support semantic cybersecurity modelling, knowledge graph construction, ontology-based threat analysis, and security assessment of EV charging infrastructure. They are intended to facilitate reproducibility and further research in EVCS cybersecurity and intelligent threat detection.

DOI
10.5281/zenodo.24202597 ↗
Autoren
Talea Huraysi
Zeitschrift / Publikationsort
Zenodo (CERN European Organization for Nuclear Research)
Publikationstyp
dataset
Status
Veröffentlicht
Veröffentlichungsdatum
13. November 2026
Cybersicherheit92% relevantOriginal lesen ↗
arXiv
VorabveröffentlichungarXiv · Computer Security·11.11.2026

Low Probability of Target Detection in Wireless Sensing by Harnessing Reconfigurable-antenna based Reflectors

In this paper, we introduce low probability of target detection (LPTD) as a new security objective for wireless sensing systems, where the existence of a target is concealed from passive wardens. Unlike conventional communication security and covert communication, LPTD aims to protect the sensed target itself from malicious detection, rather than concealing transmitted signals or communication activities. Specifically, we propose a reconfigurable-antenna-based reflector-assisted target covertness scheme, where passive reflectors deployed around the target are harnessed as artificial scattering sources to confuse the warden. By adjusting their reflection coefficients, these reflectors reshape the local scattering environment and mask the target-induced echo. We then analyze the warden's optimal detection strategy and derive the false-alarm and missed-detection probabilities, followed by a closed-form expression for its minimum total detection error probability. Numerical results validate the analysis and show that sufficient cooperation among reflectors can increase the warden's minimum total detection error probability to approximately 0.94--0.96, making target detection close to random guessing.

DOI
10.48550/arXiv.2611.16103 ↗
Autoren
Yifan Zhang, Zhentian Zhang, Jantti Riku, Yichuan Wang, Xingwang Li, Tuo Wu
Zeitschrift / Publikationsort
Nicht bereitgestellt
Publikationstyp
Vorabveröffentlichung
Status
Vorabveröffentlichung
Veröffentlichungsdatum
11. November 2026
Veröffentlichte Version
In den Metadaten der Quelle nicht verknüpft
Cybersicherheit92% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·10.11.2026

An Information-Theoretic Framework for Evaluating Class Imbalance and Rare-Attack Observability in IoT Security Benchmarks

IoT intrusion-detection studies often treat a benchmark as adequate once it is large, recent, and widely used.That assumption can fail.This study audits five public IoT and IIoT security benchmarks-CICIoT2023, Edge-IIoTset, ToN-IoT, BoT-IoT, and IoT-23-from their published class populations and introduces no synthetic classifier results; the audit combines Shannon entropy, Hill diversity, Simpson concentration, Gini inequality, rank-frequency tail fitting, rare-class observation probability, Wilson uncertainty, and a constrained power-tempered sampling model.Across the five corpora, record counts range from 461,043 to 73,370,443, normalized entropy from 0.421 to 0.600, and effective Shannon diversity from 2.20 to 3.98 classes.Size is a poor proxy for label adequacy.IoT-23 is the clearest case: its 6.05 million records include a singleton class, so a random 1% sample has only about a 1% probability of observing that label, while the worst-case Wilson half-width for a class-conditioned proportion is 0.445.We then define q i (τ) = p τ i / ∑ k p τ k and prove that its entropy is non-increasing in τ; at a 1% training budget with a target of 25 expected observations in the rarest class, the minimum-flattening exponent is 1.000 for CICIoT2023, 0.855 for Edge-IIoTset, 0.807 for ToN-IoT, 0.954 for BoT-IoT, and 0.450 for IoT-23.The last case needs an approximately 2,500-fold uplift of the singleton label.Such an uplift changes training exposure to one observation; it cannot create independent evidence.The resulting audit protocol therefore separates training rebalancing from evaluation validity and gives quantitative checks that can be applied before IoT benchmark results are interpreted as class-wide security evidence.

DOI
10.54216/jisiot.190103 ↗
Autoren
Firas Fawzy Zobary
Zeitschrift / Publikationsort
Journal of Intelligent Systems and Internet of Things
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
10. November 2026
Cybersicherheit92% relevantPDFOriginal lesen ↗
OA
FachartikelOpenAlex·09.11.2026

The Role of Deep Learning in Agriculture: A Survey on Plant Disease Detection

Plants are essential in a country's economy, climate change, and agriculture industry. Plants are susceptible to diseases brought on by bacteria, fungi, and viruses, just like people. Plant diseases have the potential to seriously impair agricultural productivity worldwide, resulting in financial losses and endangering global food security. Their early identification is crucial since it influences the growth of their respective species. In order to stop the spread of disease and lessen its impacts, early detection is essential. Manual inspection and laboratory testing are time-consuming and error prone traditional disease detection methods. Plant disease detection can now be automated, and its accuracy increased with the assist of DL, a branch of AI. Compared to traditional ML techniques, DL eliminates the need for feature engineering, so it is superior to machine learning because it does not require domain expertise. This review aims to evaluate recent advancements in Deep Learning–based approaches for plant disease detection, identify their strengths and limitations compared to traditional methods, and propose future research directions to enhance accuracy, scalability, and real-world applicability.

DOI
10.1201/9781042029907-53 ↗
Autoren
Ayushi Pandey, Suraiya Praveen, Parul Agarwal
Zeitschrift / Publikationsort
CRC Press eBooks
Publikationstyp
book chapter
Status
Veröffentlicht
Veröffentlichungsdatum
9. November 2026
Cybersicherheit92% relevantOriginal lesen ↗
OA
FachartikelOpenAlex·09.11.2026

AI Based Threat Detection System In Network

Network intrusion detection must adapt to changing traffic while preserving recognition of previously observed attacks. This paper studies the trade-off between retraining frequency, labelled sample size, and detection performance, and designs an evidence-guided security operations centre (SOC) workflow around the detector. A 100-tree Random Forest is evaluated on six constructed traffic windows with increasing proportions of DDoS and PortScan attacks. Across five paired random seeds, an output-based adaptive policy achieves 92.008% pooled incoming attack recall, compared with 92.302% for retraining every two windows, while reducing mean retraining counts from 3.0 to 1.8. No drift alarms occur in the tested stable controls. However, exact-feature grouped evaluation exposes a concentrated retention failure: one seed loses detection of 883 previously recognised old attack rows after an update. A separate validation-gate experiment accepts every candidate and therefore does not establish a protective benefit. A small FTP capture further shows that successful feature conversion does not guarantee attack recognition: all 30 flows are classified benign despite evidence of repeated failed logins. The proposed architecture connects detection with packet investigation, indicator extraction, timelines, evidence-grounded analysis, ATT&CK mapping, and analyst-approved response. Results support a limited offline cost–recall trade-off, while identifying retention, labelling realism, and extractor compatibility as unresolved requirements.

DOI
10.5281/zenodo.24077812 ↗
Autoren
Saquib Azam Khan, Shahid Farooq Malik, Shaikh Aslam Sami, Dr. Jagadishwari V
Zeitschrift / Publikationsort
Zenodo (CERN European Organization for Nuclear Research)
Publikationstyp
Zeitschriftenartikel
Status
Veröffentlicht
Veröffentlichungsdatum
9. November 2026
Cybersicherheit92% relevantOriginal lesen ↗
Erstellt mit ResearchRadarErgebnisse abgerufen: 16.11.2026, 19:01 UTCWird ungefähr alle 15 Minuten aktualisiert
Ein echtes Radar, ganz unverbindlich.Öffne jede Originalpublikation und erkunde die sortierten Ergebnisse. Ein Konto brauchst du nur zum Speichern, Anpassen oder für den Tagesüberblick.

Möchtest du dieses Radar übernehmen?

Erstelle ein kostenloses Konto mit vorbereitetem Thema, Suchbegriff, Zeitraum und Quellensatz.

Andere Vorlage testen