HIDM: Hybrid Intrusion Detection Model for Industry 4.0 Networks Using an Optimized CNN-LSTM with Transfer Learning
Abstract
:1. Introduction
- This research presents a hybrid intrusion detection model using an optimized CNN and transfer learning for cyber threat detection in Industry 4.0.
- In the proposed model, the parameters of the CNN model were optimized using the GWO method, which fine-tunes the CNN parameters, i.e., pooling size, kernel size, number of filters, number of epochs, and batch size, which helps to enhance the model’s prediction accuracy.
- Using multi-class classification, the proposed hybrid model, existing OCNN-LSTM [1], and the CNN and LSTM model were tested on two popular IIoT datasets, ToN-IoT and UNW-NB15.
- An experimental analysis was performed using various performance-measuring parameters, i.e., precision, F-measure, recall, accuracy, and detection rate; a comparison analysis was performed betweenthe existing OCNN-LSTM model and the proposed HIDM model.
- In the experimental results, the proposed HIDM model achieved a precision of 92.7% for the ToN-IoT dataset and 94.25% for the UNW-NB15 dataset.
2. Related Work
3. Materials and Methods
3.1. Dataset Details
3.1.1. UNW-NB15
3.1.2. ToN-IoT
3.2. Proposed HIDM
3.2.1. Optimized CNN usingthe GWO Method
Algorithm 1: GWO method for hyperparameter optimization for CNN |
Input: input population, Batch size Bc, Agents Ag, dimension Dn, Hyper-parameters H1, H2, H3, and H4, n is the number of iterations, Hyperparameter function Hfd, Initialization of population P1, P2 and P3 Output: Optimized hyper-parameter |
Select a data sample for training from batch input For the number of optimization, n For search desire agent Determine the optimized fitness function Select the best search agents A1, A2, and A3. Update the position of each agent by end for Update P1, P2 and P3 End Sigmoid function (H1, H2, H3,H4, Hfd) End |
3.2.2. OCNN-LSTM with Transfer Learning
3.3. Performance-Measuring Parameters
4. Experimental Results
4.1. Dataset and Data Pre-Processing
4.2. Experimental Details
4.2.1. Experiment1 for TON-IoT Dataset
4.2.2. Experiment2 for UNW-NB15 Dataset
5. Discussion and Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
Abbreviations
Abbreviation | Details |
DL | Deep learning |
TL | Transfer learning |
IDS | Intrusion detection system |
CNN | Convolution neural network |
LSTM | Long short-term memory |
DoS | Denial of service attack |
DDoS | Distributed denial of service attack |
IoT | Internet of Things |
IIoT | Industrial Internet of Things |
OCNN | Optimized CNN |
GWO | Grey wolf optimizer |
HIDM | Hybrid intrusion detection model |
H-M | Human-to-machine |
M-M | Machine-to-machine |
ML | Machine learning |
ANN | Artificial neural network |
GA | Genetic algorithm |
SVM | Support vector machine |
PSO | Particle swarm optimization |
Light-GB | Light gradient boosting |
IoD | Internet of Data |
IoP | Internet of People |
IoS | Internet of Services |
References
- Abid, H.; Javaid, M.; Singh, R.P.; Rab, S.; Suman, R. Perspectives of cybersecurity for ameliorative Industry 4.0 era: A review-based framework. Ind. Robot. Int. J. Robot. Res. Appl. 2022. online ahead of print. [Google Scholar]
- Muhammad, A.; Sadiq, S.; Alluhaidan, A.S.; Umer, M.; Ullah, S.; Nappi, M. Industry 4.0 technologies and their applications in fighting COVID-19 pandemic using deep learning techniques. Comput. Biol. Med. 2022, 145, 105418. [Google Scholar]
- Valentin, M.; Sondi, P.; Ramat, E. A review of cybersecurity guidelines for manufacturing factories in industry 4.0. IEEE Access 2021, 9, 23235–23263. [Google Scholar]
- Fraga-Lamas, P.; Lopes, S.I.; Tiago; Fernández-Caramés, M. Emerging Paradigms and Architectures for Industry 4.0 Applications. Appl. Sci. 2022, 12, 9546. [Google Scholar] [CrossRef]
- Kumar, A.A.; Karne, R.K. IIoT-IDS Network using Inception CNN Model. J. Trends Comput. Sci. Smart Technol. 2022, 4, 126–138. [Google Scholar] [CrossRef]
- Varghese, S.; Remya, S. Dog Breed Classification Using CNN. In Security Issues and Privacy Concerns in Industry 4.0 Applications; Wiley Online Library: Hoboken, NJ, USA, 2021; pp. 195–205. [Google Scholar]
- Sawik, T. A linear model for optimal cybersecurity investment in Industry 4.0 supply chains. Int. J. Prod. Res. 2022, 60, 1368–1385. [Google Scholar] [CrossRef]
- Holger, F.; Schriegel, S.; Jasperneite, J.; Trsek, H.; Adamczyk, H. Analysis of the Cyber-Security of industry 4.0 technologies based on RAMI 4.0 and identification of requirements. In Proceedings of the 2016 IEEE 21st International Conference on Emerging Technologies and Factory Automation (ETFA), Berlin, Germany, 1–6 September 2016; pp. 1–4. [Google Scholar]
- Fernández-Caramés, T.M.; Fraga-Lamas, P. Use case-based blended teaching of IIoT cybersecurity in the industry 4.0 era. Appl. Sci. 2017, 10, 5607. [Google Scholar] [CrossRef]
- Sklyar, V.; Kharchenko, V. ENISA documents in cybersecurity assurance for industry 4.0: IIoT threats and attacks scenarios. In Proceedings of the 2019 10th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS), Metz, France, 18–21 September 2019; Volume 2, pp. 1046–1049. [Google Scholar]
- Minh-Quang, T.; Elsisi, M.; Mahmoud, K.; Liu, M.K.; Lehtonen, M.; Darwish, M.M.F. Experimental setup for online fault diagnosis of induction machines via promising IoT and machine learning: Towards industry 4.0 empowerment. IEEE Access 2021, 9, 115429–115441. [Google Scholar]
- Umit, H.; Sevgilioglu, G. The evolving role of automated systems and its cyber-security issue for global business operations in Industry 4.0. Int. J. Bus. Ecosyst. Strategy 2019, 1, 1–11. [Google Scholar]
- Sumeyye, C.F.; Karakose, M. Comparative Analysis of Cyber Security Approaches Using Machine Learning in Industry 4.0. In Proceedings of the 2020 IEEE International Symposium on Systems Engineering (ISSE), Virtual, 12 October–12 November 2020; pp. 1–5. [Google Scholar]
- Kebande, V.R. Industrial Internet of things (IIoT) forensics: The forgotten concept in the race towards industry 4.0. Forensic Sci. Int. Rep. 2022, 5, 100257. [Google Scholar] [CrossRef]
- Jiewu, L.; Ye, S.; Zhou, M.; Zhao, J.L.; Liu, Q.; Guo, W.; Cao, W.; Fu, L. Blockchain-secured smart manufacturing in industry 4.0: A survey. IEEE Trans. Syst. ManCybern. Syst. 2020, 51, 237–252. [Google Scholar]
- Dimitrios, T.; Maniatakos, M. Open platform systems under scrutiny: A cybersecurity analysis of the device tree. In Proceedings of the 2018 25th IEEE International Conference on Electronics, Circuits and Systems (ICECS), Bordeaux, France, 9–12 December 2018; pp. 477–480. [Google Scholar]
- Laković, L.; Ognjanović, I.; Šendelj, R.; Injac, O. Semantically enhanced cyber security model for industry 4.0: Methodological framework. In Proceedings of the 2021 25th International Conference on Information Technology (IT), Zabljak, Montenegro, 16–20 February 2021; pp. 1–4. [Google Scholar]
- Wahid, A.; Breslin, J.G.; Intizar, M.A. Prediction of Machine Failure in Industry 4.0: A Hybrid OCNN-LSTM Framework. Appl. Sci. 2022, 12, 4221. [Google Scholar] [CrossRef]
- Yibo, H.; Zhang, D.; Cao, G.; Pan, Q. Network data analysis and anomaly detection using CNN technique for industrial control systems security. In Proceedings of the 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC), Bari, Italy, 6–9 October 2019; pp. 593–597. [Google Scholar]
- Miškuf, M.; Zolotová, I. Comparison between multi-class classifiers and deep learning with focus on industry 4.0. In Proceedings of the 2016 Cybernetics & Informatics (K&I), Levoca, Slovakia, 2–5 February 2016; pp. 1–5. [Google Scholar]
- Marianna, L.; Lazoi, M.; Corallo, A. Cybersecurity for Industry 4.0 in the current literature: A reference framework. Comput. Ind. 2018, 103, 97–110. [Google Scholar]
- Giovanna, C.; Fattori, F.; Podrecca, M.; Sartor, M. Addressing industry 4.0 cybersecurity challenges. IEEE Eng. Manag. Rev. 2019, 47, 79–86. [Google Scholar]
- Ahmed, R.S.; Ahmed, E.S.A.; Saeed, R.A. Machine learning in cyber-physical systems in industry 4.0. In Artificial Intelligence Paradigms for Smart Cyber-Physical Systems; IGI Global: Hershey, PN, USA, 2021; pp. 20–41. [Google Scholar]
- Elsisi, M.; Tran, M.Q.; Mahmoud, K.; Lehtonen, M.; Darwish, M.M. Deep learning-based industry 4.0 and internet of things towards effective energy management for smart buildings. Sensors 2021, 21, 1038. [Google Scholar] [CrossRef]
- Angelo, C.; Lazoi, M.; Lezzi, M. Cybersecurity in the context of industry 4.0: A structured classification of critical assets and business impacts. Comput. Ind. 2020, 114, 103165. [Google Scholar]
- Kumar, L.U.; Simaiya, S.; Kaur, A.; Prasad, D.; Khurana, M.; Verma, D.K.; Hassan, A. Impact of deep learning and machine learning in Industry 4.0: Impact of deep learning. In Cyber-Physical, IoT, and Autonomous Systems in Industry 4.0; CRC Press: Boca Raton, FL, USA, 2021; pp. 179–197. [Google Scholar]
- Xihua, Z.; Goyal, S.B.; Tesfayohanis, M.; Verma, C. Blockchain-Based Privacy-Preserving Approach Using SVML for Encrypted Smart City Data in the Era of IR 4.0. J. Nanomater. 2022, 2022, 7463513. [Google Scholar] [CrossRef]
- Kumar, L.U.; Simaiya, S.; Sandhu, J.K.; Trivedi, N.K.; Garg, A.; Moudgil, A. Deep Learning-Based Predictive Model for Defect Detection and Classification in Industry 4.0. In Proceedings of the 2022 International Conference on Emerging Smart Computing and Informatics (ESCI), Pune, India, 9–11 March 2022; pp. 1–5. [Google Scholar]
- Khan, R.U.; Zhang, X.; Alazab, M.; Kumar, R. An improved convolutional neural network model for intrusion detection in networks. In Proceedings of the 2019 Cybersecurity and Cyber Forensics Conference (CCC), Melbourne, Australia, 8–9 May 2019; pp. 74–77. [Google Scholar]
- Ferrag, M.A.; Maglaras, L.; Moschoyiannis, S.; Janicke, H. Deep learning for cyber security intrusion detection: Approaches, datasets, and comparative study. J. Inf. Secur. Appl. 2020, 50, 102419. [Google Scholar] [CrossRef]
- Hui, W.; Cao, Z.; Hong, B. A network intrusion detection system based on convolutional neural network. J. Intell. Fuzzy Syst. 2020, 38, 7623–7637. [Google Scholar]
- Rathore, M.S.; Poongodi, M.; Saurabh, P.; Lilhore, U.K.; Bourouis, S.; Alhakami, W.; Hamdi, M. A novel trust-based security and privacy model for Internet of Vehicles using encryption and steganography. Comput. Electr. Eng. 2022, 102, 108205. [Google Scholar] [CrossRef]
- Gupta, S.; Iyer, S.; Agarwal, G.; Manoharan, P.; Algarni, A.D.; Aldehim, G.; Raahemifar, K. Efficient Prioritization and Processor Selection Schemes for HEFT Algorithm: A Makespan Optimizer for Task Scheduling in Cloud Environment. Electronics 2022, 11, 2557. [Google Scholar] [CrossRef]
- Balyan, A.K.; Ahuja, S.; Lilhore, U.K.; Sharma, S.K.; Manoharan, P.; Algarni, A.D.; Raahemifar, K. A Hybrid Intrusion Detection Model Using EGA-PSO and Improved Random Forest Method. Sensors 2022, 22, 5986. [Google Scholar] [CrossRef] [PubMed]
- Poongodi, M.; Bourouis, S.; Ahmed, A.N.; Vijayaragavan, M.; Venkatesan KG, S.; Alhakami, W.; Hamdi, M. A novel secured multi-access edge computing based vanet with neuro-fuzzy systems based blockchain framework. Comput. Commun. 2022, 192, 48–56. [Google Scholar]
- Manoharan, P.; Walia, R.; Iwendi, C.; Ahanger, T.A.; Suganthi, S.T.; Kamruzzaman, M.M.; Hamdi, M. SVM-based generative adverserial networks for federated learning and edge computing attack model and outpoising. Expert Syst. 2022, 20, e13072. [Google Scholar] [CrossRef]
Reference | Model Used | Dataset | Outcome |
---|---|---|---|
[7] | Machine and deep learning algorithms | X-IIoTID | Deep learning methods achieved higher accuracy. |
[8] | Federated learning and edge devices | X-IIoTID | Binary classification achieved 97.89% accuracy. |
[9] | Deep neural network | ISOT, NSL-KDD, and X-IIoTID | DNN achieves high accuracy. |
[10] | Clustering algorithm, expectation max, and ANN | UNSW-NB15 and KDD99 | Expectation max achieves an accuracy of 81.25%. |
[11] | GA-LR (Genetic algorithm with Logistic Regression) | UNSW-NB15 | Accuracy of 89.87%. |
[12] | Light gradient boosting machine | UNSW-NB15 | Accuracy of 87.48%. |
[13] | Variation of long short-term memory (VLSTM) | UNSW-NB15, ToN-IoT | Precision 98.58%. |
[14] | Extreme learning machine (ELM)-based IDS | NSL-KDD and UNSW-NB15 | Accuracy 91.87%. |
[15] | Deep neural networks | UNSW-NB15, Kyoto, KDD-Cup99, CICIDS, NSL-KDD, and WSN-DS | The precision of 94.81%. |
[16] | ANN for binary classification | UNSW-NB15 | Accuracy 87.29%. |
[17] | Two-stage TABU search (TS-TS) algorithm | UNSW-NB15 | Precision 84.23%. |
[18] | Xgboost with linear regression | UNSW-NB15 and ToN-IoT | Accuracy 79.59%. |
[19] | LTSM with RNN | UNSW-NB15 | Accuracy 88.31%. |
[20] | Deep auto-encoder with LSTM | UNSW-NB15 | Accuracy 91.81%. |
[21] | Machine learning | ToN-IoT and X-IIoTID | Precision 87.84%. |
Hybrid Model | Optimized CNN with transfer learning | UNW-NB15 and ToN-IoT | 92.7%precision for ToN-IoT and 94.2% for the UNW-NB15 dataset. |
Data Based on Events | Total Records |
---|---|
Normal | 38,500 |
Back-door | 950 |
Worms | 174 |
Reconnaissance | 978 |
Fuzzers | 7500 |
DoS | 4800 |
Shellcode | 978 |
Exploits | 12,440 |
Generic | 18,554 |
Data Based on Events | Total Records |
---|---|
Normal | 79,638 |
Denial of service (DoS) | 33,753 |
Back-door | 50,811 |
Distributed denialofservice (DDoS) | 61,650 |
MITM | 105 |
Injections | 45,265 |
Ransomware | 7280 |
Scanning | 71,401 |
Cross-site scripting (XSS) | 21,089 |
Password | 17,185 |
Model Type | Parameters | Source Type | |||
---|---|---|---|---|---|
Win-7 | Win-10 | Win-10 Network | Network type | ||
OCNN-LSTM [1] | Training | 23,789 | 29,878 | 98,9145 | 1,854,791 |
Testing | 6897 | 8178 | 325,147 | 5,478,985 | |
Epoch 50 | 30 | 47 | 4 | 2 | |
OCNN-LSTM with Transfer Learning | Training | 23,789 | 29,878 | 989,145 | 1,854,791 |
Testing | 6897 | 8178 | 325,147 | 5,478,985 | |
Epoch 50 | 34 | 45 | 6 | 3 |
Model Type | Parameters | Class Type | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Normal | DoS | DDoS | Back-Door | MITM | Injections | Ransomware | Scanning | Password | XSS | Average | ||
Existing OCNN-LSTM [1] | Precision | 1 | 0.78 | 0.86 | 0.89 | 0.89 | 0.89 | 0.89 | 0.88 | 0.92 | 0.91 | 0.891 |
Recall | 1 | 0.08 | 0.07 | 0.06 | 0.91 | 0.61 | 0.45 | 0.78 | 0.35 | 0.65 | 0.496 | |
Accuracy | 1 | 0.85 | 0.87 | 0.71 | 0.91 | 0.81 | 0.89 | 0.88 | 0.94 | 0.87 | 0.873 | |
F1-Score | 1 | 0.08 | 0.07 | 0.06 | 0.91 | 0.61 | 0.45 | 0.78 | 0.25 | 0.65 | 0.486 | |
Existing CNN | Precision | 1 | 0.71 | 0.79 | 0.85 | 0.81 | 0.812 | 0.82 | 0.802 | 0.812 | 0.80 | 0.81 |
Recall | 1 | 0.06 | 0.05 | 0.05 | 0.8 | 0.55 | 0.40 | 0.70 | 0.32 | 0.60 | 0.48 | |
Accuracy | 1 | 0.81 | 0.76 | 0.63 | 0.81 | 0.71 | 0.79 | 0.78 | 0.74 | 0.77 | 0.73 | |
F1-Score | 1 | 0.07 | 0.06 | 0.05 | 0.81 | 0.54 | 0.41 | 0.71 | 0.21 | 0.61 | 0.42 | |
Existing LSTM | Precision | 1 | 0.73 | 0.81 | 0.86 | 0.84 | 0.85 | 0.86 | 0.85 | 0.87 | 0.85 | 0.85 |
Recall | 1 | 0.07 | 0.06 | 0.06 | 0.85 | 0.59 | 0.41 | 0.70 | 0.31 | 0.62 | 0.47 | |
Accuracy | 1 | 0.82 | 0.77 | 0.66 | 0.85 | 0.76 | 0.81 | 0.82 | 0.79 | 0.79 | 0.78 | |
F1-Score | 1 | 0.071 | 0.062 | 0.050 | 0.82 | 0.52 | 0.40 | 0.71 | 0.21 | 0.60 | 0.44 | |
Proposed OCNN-LSTM with Transfer Learning | Precision | 1 | 0.91 | 0.9 | 0.925 | 0.932 | 0.918 | 0.91 | 0.89 | 0.93 | 0.95 | 0.927 |
Recall | 1 | 0.081 | 0.074 | 0.065 | 0.94 | 0.689 | 0.51 | 0.79 | 0.41 | 0.68 | 0.5239 | |
Accuracy | 1 | 0.97 | 0.98 | 0.91 | 0.93 | 0.96 | 0.94 | 0.89 | 0.95 | 0.91 | 0.944 | |
F1-Score | 1 | 0.091 | 0.087 | 0.098 | 0.94 | 0.78 | 0.56 | 0.87 | 0.45 | 0.78 | 0.566 |
Model Type | Parameters | Class Type | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Normal | Back-Door | Worms | Reconnaissance | Fuzzers | DoS | ShellCode | Exploits | Generic | Average | ||
Existing OCNN-LSTM [1] | Precision | 1 | 0.87 | 0.89 | 0.91 | 0.92 | 0.91 | 0.901 | 0.867 | 0.914 | 0.909 |
Recall | 1 | 0.078 | 0.089 | 0.078 | 0.95 | 0.78 | 0.55 | 0.67 | 0.45 | 0.516 | |
Accuracy | 1 | 0.89 | 0.87 | 0.91 | 0.9 | 0.87 | 0.914 | 0.89 | 0.95 | 0.910 | |
F1-Score | 1 | 0.66 | 0.45 | 0.55 | 0.087 | 0.074 | 0.056 | 0.84 | 0.45 | 0.463 | |
OCNN-LSTM with Transfer Learning | Precision | 1 | 0.91 | 0.94 | 0.93 | 0.94 | 0.93 | 0.95 | 0.93 | 0.95 | 0.942 |
Recall | 1 | 0.085 | 0.0901 | 0.0845 | 0.96 | 0.812 | 0.65 | 0.712 | 0.556 | 0.550 | |
Accuracy | 1 | 0.91 | 0.86 | 0.92 | 0.93 | 0.914 | 0.923 | 0.923 | 0.96 | 0.927 | |
F1-Score | 1 | 0.067 | 0.54 | 0.65 | 0.086 | 0.087 | 0.066 | 0.89 | 0.67 | 0.451 | |
Existing CNN | Precision | 1 | 0.81 | 0.82 | 0.82 | 0.82 | 0.81 | 0.801 | 0.827 | 0.814 | 0.809 |
Recall | 1 | 0.068 | 0.079 | 0.068 | 0.85 | 0.68 | 0.45 | 0.57 | 0.41 | 0.456 | |
Accuracy | 1 | 0.809 | 0.807 | 0.801 | 0.81 | 0.81 | 0.814 | 0.809 | 0.82 | 0.810 | |
F1-Score | 1 | 0.566 | 0.40 | 0.45 | 0.077 | 0.070 | 0.046 | 0.74 | 0.35 | 0.413 | |
Existing LSTM | Precision | 1 | 0.84 | 0.83 | 0.84 | 0.83 | 0.83 | 0.821 | 0.847 | 0.824 | 0.829 |
Recall | 1 | 0.071 | 0.075 | 0.071 | 0.88 | 0.72 | 0.52 | 0.59 | 0.46 | 0.48 | |
Accuracy | 1 | 0.829 | 0.827 | 0.831 | 0.831 | 0.841 | 0.854 | 0.839 | 0.852 | 0.850 | |
F1-Score | 1 | 0.596 | 0.46 | 0.465 | 0.087 | 0.080 | 0.056 | 0.84 | 0.45 | 0.483 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Lilhore, U.K.; Manoharan, P.; Simaiya, S.; Alroobaea, R.; Alsafyani, M.; Baqasah, A.M.; Dalal, S.; Sharma, A.; Raahemifar, K. HIDM: Hybrid Intrusion Detection Model for Industry 4.0 Networks Using an Optimized CNN-LSTM with Transfer Learning. Sensors 2023, 23, 7856. https://doi.org/10.3390/s23187856
Lilhore UK, Manoharan P, Simaiya S, Alroobaea R, Alsafyani M, Baqasah AM, Dalal S, Sharma A, Raahemifar K. HIDM: Hybrid Intrusion Detection Model for Industry 4.0 Networks Using an Optimized CNN-LSTM with Transfer Learning. Sensors. 2023; 23(18):7856. https://doi.org/10.3390/s23187856
Chicago/Turabian StyleLilhore, Umesh Kumar, Poongodi Manoharan, Sarita Simaiya, Roobaea Alroobaea, Majed Alsafyani, Abdullah M. Baqasah, Surjeet Dalal, Ashish Sharma, and Kaamran Raahemifar. 2023. "HIDM: Hybrid Intrusion Detection Model for Industry 4.0 Networks Using an Optimized CNN-LSTM with Transfer Learning" Sensors 23, no. 18: 7856. https://doi.org/10.3390/s23187856
APA StyleLilhore, U. K., Manoharan, P., Simaiya, S., Alroobaea, R., Alsafyani, M., Baqasah, A. M., Dalal, S., Sharma, A., & Raahemifar, K. (2023). HIDM: Hybrid Intrusion Detection Model for Industry 4.0 Networks Using an Optimized CNN-LSTM with Transfer Learning. Sensors, 23(18), 7856. https://doi.org/10.3390/s23187856