Big Data Empowered Cyber Security Intelligence via GhostLSTM model and Adaptive Entropy based Attack Mitigation

In recent years, the implementation of Big Data analytics has risen across multiple areas such as transportation, communications and healthcare. Intrusion Detection Systems play a crucial role in ensuring network security by recognizing and mitigating cyberattacks. This paper proposes a cyber-attack detection framework from a big data perspective via an IGN-LSTM model, along with attack mitigation using an enhanced entropy (Eent) method. The proposed framework begins with a preprocessing phase, where the input unbalanced data is initially balanced, followed by the min-max normalization technique. Then, the normalized data is processed using a MapReduce framework. According to the mapper phase, features like raw data, higher-order statistical features and Renyi entropy-based features are extracted. In the reducer phase, all mapper results are aggregated, which produces the final extracted feature set. After that, feature selection is performed using the Self-improved Wolverine Optimization Algorithm (Si-WlOA), which enhances population diversity, global convergence speed and overall optimization efficiency. The selected features are input into a hybrid (IGN-LSTM) classification model combining the improved GhostNet (IGN) and LSTM models, where context-aware soft selection dynamically adjusts feature focus to optimize classification performance. Lastly, the attacks are mitigated through the Enhanced Entropy (Eent) process, which can detect small deviations in the data, which identified attack nodes effectively in the network. Moreover, experimental results prove that the proposed framework outperforms conventional approaches in terms of detection and mitigation performance. The proposed IGN-LSTM model has achieved better accuracy by 96.5% at 80% and 97.2% at 90%.

Authors

Publication Details

Journal
Discrete Mathematics Algorithms and Applications
Published
2026-11-19
DOI
https://doi.org/10.1142/s1793830926501181
Primary Topic
Network Security and Intrusion Detection
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Big Data Empowered Cyber Security Intelligence via GhostLSTM model and Adaptive Entropy based Attack Mitigation

Kummari Dharmendra, Palle Vijaya Kumari
Discrete Mathematics Algorithms and Applications
Network Security and Intrusion Detection
article

Big Data Empowered Cyber Security Intelligence via GhostLSTM model and Adaptive Entropy based Attack Mitigation

Kummari Dharmendra, Palle Vijaya Kumari
article en

Abstract

In recent years, the implementation of Big Data analytics has risen across multiple areas such as transportation, communications and healthcare. Intrusion Detection Systems play a crucial role in ensuring network security by recognizing and mitigating cyberattacks. This paper proposes a cyber-attack detection framework from a big data perspective via an IGN-LSTM model, along with attack mitigation using an enhanced entropy (Eent) method. The proposed framework begins with a preprocessing phase, where the input unbalanced data is initially balanced, followed by the min-max normalization technique. Then, the normalized data is processed using a MapReduce framework. According to the mapper phase, features like raw data, higher-order statistical features and Renyi entropy-based features are extracted. In the reducer phase, all mapper results are aggregated, which produces the final extracted feature set. After that, feature selection is performed using the Self-improved Wolverine Optimization Algorithm (Si-WlOA), which enhances population diversity, global convergence speed and overall optimization efficiency. The selected features are input into a hybrid (IGN-LSTM) classification model combining the improved GhostNet (IGN) and LSTM models, where context-aware soft selection dynamically adjusts feature focus to optimize classification performance. Lastly, the attacks are mitigated through the Enhanced Entropy (Eent) process, which can detect small deviations in the data, which identified attack nodes effectively in the network. Moreover, experimental results prove that the proposed framework outperforms conventional approaches in terms of detection and mitigation performance. The proposed IGN-LSTM model has achieved better accuracy by 96.5% at 80% and 97.2% at 90%.

Discrete Mathematics Algorithms and Applications
Network Security and Intrusion Detection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon