Chandrakanth Reddy Borra
Research Subjects
Chandrakanth Reddy Borra has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
Ramya Vani Rayala
University of the Cumberlands,United States
17 shared publicationsSrinivas Cheekati
University of the Cumberlands,United States
17 shared publicationsPiyush Kumar Pareek
Nitte Meenakshi Institute of Technology,Department of AIML,Bengaluru
7 shared publicationsVani Vasudevan
Nitte Meenakshi Institute of Technology Nitte (Deemed to be University),Department of Computer Science Engineering,Bengaluru,India
4 shared publicationsRuchita Singhania
Dayananda Sagar Academy of Technology and Management,Department of AIML,Bangaluru,India
2 shared publicationsZokir Mamadiyarov
Mamun University,Department of Economics,Khiva,Uzbekistan
2 shared publications
Latest Publications
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-9
Enhanced DDoS Attack Detection Using Shepard Interpolation Neural Network with Artificial Rabbit Optimizer and SMOTE Balancing on Kaggle Dataset
Strong defence mechanisms are necessary to defend the availability and integrity of network infrastructure in the ever-changing cybersecurity landscape, where Distributed Denial of Service (DDoS) attacks are becoming more common. The capacity of Deep Learning (DL) models to automatically learn feature representations and discern complicated patterns within network traffic data has made them a potential strategy for DDoS attack finding and mitigation. The design of a critical cybersecurity threat that disrupts network operations and causes considerable economic losses internationally also affects how well DL models fight against developing attacks. The Shepard Interpolation Neural Network (SINN) classifier, fine-tuned with the Artificial Rabbit Optimiser (ARO), is proposed as a new method for DDoS assault detection in this paper. The technique takes advantage of the Kaggle DDoS dataset, which is naturally skewed and poses problems for conventional ML models. The solution is to use the Synthetic Minority Oversampling Technique (SMOTE) to level the playing field in the dataset, which will result in better model presentation in every class. Thanks to ARO’s hyperparameter optimisation, the SINN classifier—which is already well-known for its interpolation capabilities and resilience when dealing with nonlinear patterns—achieves even better detection accuracy. Experiments show that the suggested model outdoes state-of-the-art machine learning tactics for DDoS attack finding in terms of accuracy, recall, and F1-score. Reliable classification results are contributed to by the incorporation of SMOTE, which successfully mitigates bias induced by class imbalance. In addition to establishing a scalable framework for practical applications, this work showcases SINN-ARO’s capabilities in protecting networks from ever-changing cyber threats. Investigating how well this approach works in real-time, dynamic network settings and expanding its applicability to identify different forms of cyberattacks are the primary goals of future studies.
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-9
Enhanced DDoS Attack Detection Using Shepard Interpolation Neural Network with Artificial Rabbit Optimizer and SMOTE Balancing on Kaggle Dataset
Strong defence mechanisms are necessary to defend the availability and integrity of network infrastructure in the ever-changing cybersecurity landscape, where Distributed Denial of Service (DDoS) attacks are becoming more common. The capacity of Deep Learning (DL) models to automatically learn feature representations and discern complicated patterns within network traffic data has made them a potential strategy for DDoS attack finding and mitigation. The design of a critical cybersecurity threat that disrupts network operations and causes considerable economic losses internationally also affects how well DL models fight against developing attacks. The Shepard Interpolation Neural Network (SINN) classifier, fine-tuned with the Artificial Rabbit Optimiser (ARO), is proposed as a new method for DDoS assault detection in this paper. The technique takes advantage of the Kaggle DDoS dataset, which is naturally skewed and poses problems for conventional ML models. The solution is to use the Synthetic Minority Oversampling Technique (SMOTE) to level the playing field in the dataset, which will result in better model presentation in every class. Thanks to ARO’s hyperparameter optimisation, the SINN classifier—which is already well-known for its interpolation capabilities and resilience when dealing with nonlinear patterns—achieves even better detection accuracy. Experiments show that the suggested model outdoes state-of-the-art machine learning tactics for DDoS attack finding in terms of accuracy, recall, and F1-score. Reliable classification results are contributed to by the incorporation of SMOTE, which successfully mitigates bias induced by class imbalance. In addition to establishing a scalable framework for practical applications, this work showcases SINN-ARO’s capabilities in protecting networks from ever-changing cyber threats. Investigating how well this approach works in real-time, dynamic network settings and expanding its applicability to identify different forms of cyberattacks are the primary goals of future studies.
- Open Access
- 1 January 2026
- Published by Elsevier in SSRN Electronic Journal
Quantitative Correlational Analysis of Factors Affecting Privacy Perceptions towards AI Devices
- 1 January 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessQuantitative Correlational Analysis of Factors Affecting Privacy Perceptions towards AI Devices
- Open Access
- 31 December 2025
- Published by International Information and Engineering Technology Association in Traitement du Signal
- Vol. 42 (6) , 3045-3060
Artificial Intelligence-Based Facial Emotion Valence Detection Model Using Advanced Feature Engineering Techniques
- 31 December 2025
- Published by International Information and Engineering Technology Association in Traitement du Signal
- Vol. 42 (6) , 3045-3060
Open AccessArtificial Intelligence-Based Facial Emotion Valence Detection Model Using Advanced Feature Engineering Techniques
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Optimized Deep Learning for Diabetes Detection: A BGRU-based Approach with SA-GSO Hyperparameter Tuning
Every year, millions of people are impacted by diabetes, a metabolic disease. A lower quality of life and an increased risk of critical organ failure are both linked to it. Diabetes management relies on early detection and consistent monitoring. Modern technological advancements in remote patient monitoring have the potential to revolutionise intervention and therapy paradigms. Using the PIMA Indian Diabetes dataset, this work applies machine learning approaches to recover the prediction accuracy of diabetes diagnosis. The dataset was subjected to thorough preparation, which involved imputation of Multivariate Imputation by Chained Equations (MICE) technique and feature scaling standardisation. The most important attributes leading to diabetes prediction were efficiently discovered using the student psychology-based optimisation (SPBO) method, which was used for feature selection. A BGRU model was used for disease classification, which successfully captured sequential dependencies in patient data and reduced the problems associated with gradients that are common in traditional recurrent neural networks. To further enhance model efficiency and diagnostic accuracy, the SA-GSO method was used to optimise hyperparameter tuning. Diabetes prediction can be automated and improved with the help of the suggested method, which provides a solid foundation for medical applications. In order to make the dataset more applicable in clinical practice, further study will involve investigating different imputation approaches, reviewing more feature selection methodologies, and increasing the dataset's diversity.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Optimized Deep Learning for Diabetes Detection: A BGRU-based Approach with SA-GSO Hyperparameter Tuning
Every year, millions of people are impacted by diabetes, a metabolic disease. A lower quality of life and an increased risk of critical organ failure are both linked to it. Diabetes management relies on early detection and consistent monitoring. Modern technological advancements in remote patient monitoring have the potential to revolutionise intervention and therapy paradigms. Using the PIMA Indian Diabetes dataset, this work applies machine learning approaches to recover the prediction accuracy of diabetes diagnosis. The dataset was subjected to thorough preparation, which involved imputation of Multivariate Imputation by Chained Equations (MICE) technique and feature scaling standardisation. The most important attributes leading to diabetes prediction were efficiently discovered using the student psychology-based optimisation (SPBO) method, which was used for feature selection. A BGRU model was used for disease classification, which successfully captured sequential dependencies in patient data and reduced the problems associated with gradients that are common in traditional recurrent neural networks. To further enhance model efficiency and diagnostic accuracy, the SA-GSO method was used to optimise hyperparameter tuning. Diabetes prediction can be automated and improved with the help of the suggested method, which provides a solid foundation for medical applications. In order to make the dataset more applicable in clinical practice, further study will involve investigating different imputation approaches, reviewing more feature selection methodologies, and increasing the dataset's diversity.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Lightweight Deep Learning Framework for Human Activity Recognition: A MobileNet-LSTM Approach with WHOA Optimization
Medical informatics, HCI, surveillance, and task monitoring systems use computer vision for human activity classification in novel ways. To aid doctors in medication reactions and diagnosis, patients' activities must be analyzed and categorized. Machine learning and soft computational algorithms are used for human activity recognition from movies and photographs, but more advanced computer vision approaches are being investigated. This study proposed automated HAR classification using MobileNet V2 and LSTM. MobileNet V2 beat its predecessors in efficiency and accuracy testing on low-power computing devices. For different action recognition scenarios, this study uses RGB, depth, and skeletal joint data from UTKinect-Action3D and CMU Mocap3D. A systematic preprocessing pipeline includes frame segmentation, temporal sampling, optical flow computation, Gaussian blur, color jittering, and random flips to improve model performance. Interpolation and mean imputation provide dataset completeness, while quality control procedures improve dataset dependability. MobileNet, intended for low-computation contexts, captures action sequence temporal dependencies with LSTM. Fine-tuning model hyperparameters with the Wildebeest Herd Optimization Algorithm (WHOA) improves efficiency and accuracy. Experimental results show the proposed model can achieve excellent recognition performance while being computationally efficient, making it suitable for real-time HAR applications.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Lightweight Deep Learning Framework for Human Activity Recognition: A MobileNet-LSTM Approach with WHOA Optimization
Medical informatics, HCI, surveillance, and task monitoring systems use computer vision for human activity classification in novel ways. To aid doctors in medication reactions and diagnosis, patients' activities must be analyzed and categorized. Machine learning and soft computational algorithms are used for human activity recognition from movies and photographs, but more advanced computer vision approaches are being investigated. This study proposed automated HAR classification using MobileNet V2 and LSTM. MobileNet V2 beat its predecessors in efficiency and accuracy testing on low-power computing devices. For different action recognition scenarios, this study uses RGB, depth, and skeletal joint data from UTKinect-Action3D and CMU Mocap3D. A systematic preprocessing pipeline includes frame segmentation, temporal sampling, optical flow computation, Gaussian blur, color jittering, and random flips to improve model performance. Interpolation and mean imputation provide dataset completeness, while quality control procedures improve dataset dependability. MobileNet, intended for low-computation contexts, captures action sequence temporal dependencies with LSTM. Fine-tuning model hyperparameters with the Wildebeest Herd Optimization Algorithm (WHOA) improves efficiency and accuracy. Experimental results show the proposed model can achieve excellent recognition performance while being computationally efficient, making it suitable for real-time HAR applications.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Enhancing Renewable Energy Forecasting using Roosters Optimization Algorithm and Hybrid Deep Learning Models
In recent years, there significant upsurge in the utilization of renewable resources for electricity generation. Consequently, accurate short-term predicting of renewable power production has become crucial for power system operations. However, Renewable Power Production Forecasting (RPPF) presents unique challenges due to the intermittent and uncertain nature of renewable energy sources. This study presents a deep learning-based approach for analyzing renewable energy consumption trends using a comprehensive dataset obtained from Kaggle. The dataset covers energy-related parameters from 175 countries over a 20-year period (2000–2020). To ensure data quality, preprocessing techniques such as missing data imputation, feature encoding, and Pearson correlation analysis were applied. The Roosters Optimization Algorithm (ROA) was employed for feature selection and hyperparameter tuning, optimizing computational efficiency and model performance. A hybrid Bidirectional Long-short term memory with Gated Recurrent unit (BLSTM-GRU) model was then implemented for classification, leveraging bidirectional dependencies and contextual information extraction for enhanced predictive accuracy. The proposed methodology effectively improves energy trend forecasting, demonstrating the potential of deep learning and bio-inspired optimization techniques in energy data analysis. Future work will explore additional optimization strategies and advanced neural architectures to further enhance model performance and generalizability.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Enhancing Renewable Energy Forecasting using Roosters Optimization Algorithm and Hybrid Deep Learning Models
In recent years, there significant upsurge in the utilization of renewable resources for electricity generation. Consequently, accurate short-term predicting of renewable power production has become crucial for power system operations. However, Renewable Power Production Forecasting (RPPF) presents unique challenges due to the intermittent and uncertain nature of renewable energy sources. This study presents a deep learning-based approach for analyzing renewable energy consumption trends using a comprehensive dataset obtained from Kaggle. The dataset covers energy-related parameters from 175 countries over a 20-year period (2000–2020). To ensure data quality, preprocessing techniques such as missing data imputation, feature encoding, and Pearson correlation analysis were applied. The Roosters Optimization Algorithm (ROA) was employed for feature selection and hyperparameter tuning, optimizing computational efficiency and model performance. A hybrid Bidirectional Long-short term memory with Gated Recurrent unit (BLSTM-GRU) model was then implemented for classification, leveraging bidirectional dependencies and contextual information extraction for enhanced predictive accuracy. The proposed methodology effectively improves energy trend forecasting, demonstrating the potential of deep learning and bio-inspired optimization techniques in energy data analysis. Future work will explore additional optimization strategies and advanced neural architectures to further enhance model performance and generalizability.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
AI-Driven Tomato Quality Assessment: A Deep Learning Approach for Image-based Defect Detection and Classification
Agriculture is crucial to India's economic success. Due to increased demand for premium products, fruit grading is crucial to agriculture. Fruit grading by hand is time-consuming, inaccurate, and inefficient. The automatic grading system cuts processing time and errors. Domestic and foreign markets want tomatoes. Grading tomato fruit requires caution due to its fragility. The research recommends a computer vision-based approach to automatically rate tomato fruits. A strong deep learning method for tomato picture acquisition, feature extraction, and classification is presented in this paper. Using an NVIDIA Jetson TX1 system with an RGB camera, 2400 photos were taken of 600 tomatoes of various shapes, sizes, colors, and flaws. Actual sorting mechanisms were simulated by imaging each tomato from four angles. A Skip-GRU network module bypassed extraneous information and used a self-attention method to improve feature extraction. The modified cheetah optimization algorithm (MCOA) improved feature selection for classification. Using convolutional and bidirectional long short-term memory (LSTM) networks with an effective channel attention (ECA) mechanism, an Adaptive CNN-BLSTM model extracted and prioritized significant information for classification. This method accurately detects tomato flaws and classifies them according to OECD and USDA criteria. The dataset has been made public to enhance AI-driven agriculture applications.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
AI-Driven Tomato Quality Assessment: A Deep Learning Approach for Image-based Defect Detection and Classification
Agriculture is crucial to India's economic success. Due to increased demand for premium products, fruit grading is crucial to agriculture. Fruit grading by hand is time-consuming, inaccurate, and inefficient. The automatic grading system cuts processing time and errors. Domestic and foreign markets want tomatoes. Grading tomato fruit requires caution due to its fragility. The research recommends a computer vision-based approach to automatically rate tomato fruits. A strong deep learning method for tomato picture acquisition, feature extraction, and classification is presented in this paper. Using an NVIDIA Jetson TX1 system with an RGB camera, 2400 photos were taken of 600 tomatoes of various shapes, sizes, colors, and flaws. Actual sorting mechanisms were simulated by imaging each tomato from four angles. A Skip-GRU network module bypassed extraneous information and used a self-attention method to improve feature extraction. The modified cheetah optimization algorithm (MCOA) improved feature selection for classification. Using convolutional and bidirectional long short-term memory (LSTM) networks with an effective channel attention (ECA) mechanism, an Adaptive CNN-BLSTM model extracted and prioritized significant information for classification. This method accurately detects tomato flaws and classifies them according to OECD and USDA criteria. The dataset has been made public to enhance AI-driven agriculture applications.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Intelligent Air Pollution Forecasting: A Deep Learning Approach with 1D-CNN and Kookaburra Optimization
Air quality prediction is essential for environmental monitoring and reducing air pollution's health effects. Machine learning and deep learning models like 1D-CNN can forecast future air quality by studying data trends. Air quality predictions are crucial for environmental monitoring and public health. Traditional ML models often miss complex geographical and temporal relationships in air quality data. To demonstrate a kookaburra-optimized one-dimensional convolutional neural network (1D-CNN) to improve prediction efficiency and accuracy. The 1D-CNN model extracts important properties to accurately capture time-series air quality data trends and swings. The kookaburra optimization method dynamically fine-tunes crucial model parameters for optimal network architecture and decreased computational costs. Experimental results reveal that the 1D-CNN with kookaburra optimization outperforms traditional models in accuracy, prediction error, and generalization across different air quality datasets. The findings show that deep learning and nature-inspired optimisation can be used in green computing. Future research can integrate attention processes, hybrid deep learning architectures, and real-time IoT sensor networks to improve air quality forecasts. The framework promotes intelligent air quality monitoring systems by enabling proactive pollution control and public health safety decisions.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Intelligent Air Pollution Forecasting: A Deep Learning Approach with 1D-CNN and Kookaburra Optimization
Air quality prediction is essential for environmental monitoring and reducing air pollution's health effects. Machine learning and deep learning models like 1D-CNN can forecast future air quality by studying data trends. Air quality predictions are crucial for environmental monitoring and public health. Traditional ML models often miss complex geographical and temporal relationships in air quality data. To demonstrate a kookaburra-optimized one-dimensional convolutional neural network (1D-CNN) to improve prediction efficiency and accuracy. The 1D-CNN model extracts important properties to accurately capture time-series air quality data trends and swings. The kookaburra optimization method dynamically fine-tunes crucial model parameters for optimal network architecture and decreased computational costs. Experimental results reveal that the 1D-CNN with kookaburra optimization outperforms traditional models in accuracy, prediction error, and generalization across different air quality datasets. The findings show that deep learning and nature-inspired optimisation can be used in green computing. Future research can integrate attention processes, hybrid deep learning architectures, and real-time IoT sensor networks to improve air quality forecasts. The framework promotes intelligent air quality monitoring systems by enabling proactive pollution control and public health safety decisions.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 782-788
Cybersecurity Threat Detection Using OpCyNet and DBRA: A Deep Learning Approach for DDoS Attack Mitigation on CICDDoS2019
Threatening services by overwhelming targeted systems with traffic from numerous sources, distributed denial-of-service (DDoS) assaults are a major concern for computer networks and systems. It is now an essential cybersecurity duty to detect these threats in real-time. Unfortunately, present approaches for detecting DDoS attacks have a hard time capturing the intricate patterns of attack traffic and have a high false positive rate. This research proposes an enhanced approach for detecting DDoS attacks using an Optimized Cybernet Model (OpCyNet), a deep learning framework integrated with the Developed Battle Royale Optimization Algorithm (DBRA) using CICDDoS2019 dataset. In order to extract, improve, and classify characteristics of network traffic, the OpCyNet model has eleven learnable layers. These layers include three fully connected (FC) layers and eight convolutional layers. For better generalization and to avoid overfitting, the model uses dropout layers, max-pooling, batch normalization, leaky ReLU activation, and others. Additionally, cross-channel normalization and softmax classification enhance the model's ability to distinguish between normal and malicious network traffic patterns. The DBRA metaheuristic algorithm is used to further optimize feature selection and classification, which greatly reduces computational cost while keeping detection accuracy high. Experimental results on the CICDDoS2019 dataset demonstrate that the OpCyNet-DBRA model outperforms conventional deep learning and machine learning approaches in terms of accuracy, precision, recall, F1-score, and false positive rate (FPR). When applied to real-world network settings, the suggested framework improves threat detection and mitigation while being efficient, scalable, and intelligent. To enhance cybersecurity resilience in the face of ever-changing distributed denial-of-service (DDoS) attacks, this paper proposes a method for cyberattack detection that combines deep learning with metaheuristics. To further improve the proposed system's effectiveness, future work will center on real-time deployment, adversarial attack resistance, and cross-dataset validation..
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 782-788
Cybersecurity Threat Detection Using OpCyNet and DBRA: A Deep Learning Approach for DDoS Attack Mitigation on CICDDoS2019
Threatening services by overwhelming targeted systems with traffic from numerous sources, distributed denial-of-service (DDoS) assaults are a major concern for computer networks and systems. It is now an essential cybersecurity duty to detect these threats in real-time. Unfortunately, present approaches for detecting DDoS attacks have a hard time capturing the intricate patterns of attack traffic and have a high false positive rate. This research proposes an enhanced approach for detecting DDoS attacks using an Optimized Cybernet Model (OpCyNet), a deep learning framework integrated with the Developed Battle Royale Optimization Algorithm (DBRA) using CICDDoS2019 dataset. In order to extract, improve, and classify characteristics of network traffic, the OpCyNet model has eleven learnable layers. These layers include three fully connected (FC) layers and eight convolutional layers. For better generalization and to avoid overfitting, the model uses dropout layers, max-pooling, batch normalization, leaky ReLU activation, and others. Additionally, cross-channel normalization and softmax classification enhance the model's ability to distinguish between normal and malicious network traffic patterns. The DBRA metaheuristic algorithm is used to further optimize feature selection and classification, which greatly reduces computational cost while keeping detection accuracy high. Experimental results on the CICDDoS2019 dataset demonstrate that the OpCyNet-DBRA model outperforms conventional deep learning and machine learning approaches in terms of accuracy, precision, recall, F1-score, and false positive rate (FPR). When applied to real-world network settings, the suggested framework improves threat detection and mitigation while being efficient, scalable, and intelligent. To enhance cybersecurity resilience in the face of ever-changing distributed denial-of-service (DDoS) attacks, this paper proposes a method for cyberattack detection that combines deep learning with metaheuristics. To further improve the proposed system's effectiveness, future work will center on real-time deployment, adversarial attack resistance, and cross-dataset validation..
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 774-781
Intelligent Cybersecurity for IoT: A Hybrid QRIME-SDPN Approach for Network Attack Detection on CIC-IoT-2023
The need for intelligent intrusion detection systems (IDS) has arisen due to the fact that IoT networks are becoming more vulnerable to complex cyber threats due to their fast growth. Problems with high-dimensional data, redundant characteristics, and changing assault patterns are common causes of performance constraints in traditional intrusion detection systems. This paper presents a hybrid intrusion detection system that syndicates a Stacked Deep Polynomial Network (SDPN) classifier with QRIME-based feature selection to efficiently and accurately detect network attacks on CIC-IoT-2023 dataset. The goal is to tackle these problems. In order to improve computational efficiency, the Quantum-Inspired Redundant and Irrelevant feature Minimization and Extraction (QRIME) method is used to decrease the dimensionality of features. This method ensures that only the most significant characteristics are used for classification. By using deep polynomial transformations, the SDPN classifier is able to improve its generalization capabilities and catch intricate attack patterns. Together, they enable the model to reduce false positives while efficiently detecting varied IoT-based assaults. The suggested QRIMESDPN model beats both conventional and deep learning models in terms of recall, accuracy, precision, $\mathbf{F 1}$-score, besides false alarm rate (FAR) when tested on the CIC-IoT-2023 dataset. The model's successful handling of high-dimensional network traffic data, with a balance among computational efficiency and detection capability, is demonstrated by the findings. The overall result of this study is an innovative and scalable method for intrusion detection in the Internet of Things (IoT) that combines deep polynomial-based classification with feature selection influenced by quantum mechanics. To further improve IDS performance in dynamic IoT environments, future work can centre on real-time deployment, adversarial resilience, and transfer learning. In light of the critical requirement for flexible intrusion detection systems to counter evolving cyber threats, this study makes a substantial contribution to Internet of Things security.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 774-781
Intelligent Cybersecurity for IoT: A Hybrid QRIME-SDPN Approach for Network Attack Detection on CIC-IoT-2023
The need for intelligent intrusion detection systems (IDS) has arisen due to the fact that IoT networks are becoming more vulnerable to complex cyber threats due to their fast growth. Problems with high-dimensional data, redundant characteristics, and changing assault patterns are common causes of performance constraints in traditional intrusion detection systems. This paper presents a hybrid intrusion detection system that syndicates a Stacked Deep Polynomial Network (SDPN) classifier with QRIME-based feature selection to efficiently and accurately detect network attacks on CIC-IoT-2023 dataset. The goal is to tackle these problems. In order to improve computational efficiency, the Quantum-Inspired Redundant and Irrelevant feature Minimization and Extraction (QRIME) method is used to decrease the dimensionality of features. This method ensures that only the most significant characteristics are used for classification. By using deep polynomial transformations, the SDPN classifier is able to improve its generalization capabilities and catch intricate attack patterns. Together, they enable the model to reduce false positives while efficiently detecting varied IoT-based assaults. The suggested QRIMESDPN model beats both conventional and deep learning models in terms of recall, accuracy, precision, $\mathbf{F 1}$-score, besides false alarm rate (FAR) when tested on the CIC-IoT-2023 dataset. The model's successful handling of high-dimensional network traffic data, with a balance among computational efficiency and detection capability, is demonstrated by the findings. The overall result of this study is an innovative and scalable method for intrusion detection in the Internet of Things (IoT) that combines deep polynomial-based classification with feature selection influenced by quantum mechanics. To further improve IDS performance in dynamic IoT environments, future work can centre on real-time deployment, adversarial resilience, and transfer learning. In light of the critical requirement for flexible intrusion detection systems to counter evolving cyber threats, this study makes a substantial contribution to Internet of Things security.