Scholarlink

<p>Vehicular Ad-hoc Networks (VANETs) enhance road safety by enabling vehicle-to-vehicle communication. This study aims to enhance road safety in VANETs by developing an efficient and accurate model for real-time threat classification. The primary research question investigates how a hybrid machine learning approach can improve threat detection performance in VANET-based intelligent transportation systems. Data was collected from simulated VANET environments, composed of vehicle positions, speeds, and communication metrics. Preprocessing included data cleaning and Min-Max normalization to reduce noise and standardize input features. Principal Component Analysis (PCA) was used for dimensionality reduction, and Recursive Feature Elimination (RFE) identified the most significant features. A novel hybrid classifier, Dynamic Support Vector Machine-Tree Classifier (DSVMTC), was developed and implemented in Python using the Scikit-learn library. The DSVMTC model demonstrated significant improvements in classification accuracy, response time, and computational efficiency compared to traditional methods. The model performed better than conventional and deep learning baselines, having an accuracy of 98.9%, precision of 97.8%, recall of 97.9%, F1-score of 96.9%, ROC-AUC of 0.98, detection accuracy of 98.5%, false positive rate of 1.8%, and average latency of 42 ms, illustrating applicability to real-time use. The proposed DSVMTC model effectively improves real-time threat classification in VANET-based transportation systems. The hybrid architecture not only boosts detection accuracy but also ensures low-latency processing, making it suitable for real-world intelligent transportation networks. Future research will focus on optimizing the model for large-scale deployment and real-time adaptability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Hybrid Machine Learning Approach for Real-Time Threat Classification in VANET-Based Transportation Systems

  • 期刊论文

  • 发表日期：

    2025年12月26日

查看全文
  • Mridula Gupta,
  • M. P. Karthikeyan,
  • Prabhat Kumar Sahu,
  • N. Hanuman Reddy,
  • A. Alli,
  • Yogesh Jadhav

摘要

Vehicular Ad-hoc Networks (VANETs) enhance road safety by enabling vehicle-to-vehicle communication. This study aims to enhance road safety in VANETs by developing an efficient and accurate model for real-time threat classification. The primary research question investigates how a hybrid machine learning approach can improve threat detection performance in VANET-based intelligent transportation systems. Data was collected from simulated VANET environments, composed of vehicle positions, speeds, and communication metrics. Preprocessing included data cleaning and Min-Max normalization to reduce noise and standardize input features. Principal Component Analysis (PCA) was used for dimensionality reduction, and Recursive Feature Elimination (RFE) identified the most significant features. A novel hybrid classifier, Dynamic Support Vector Machine-Tree Classifier (DSVMTC), was developed and implemented in Python using the Scikit-learn library. The DSVMTC model demonstrated significant improvements in classification accuracy, response time, and computational efficiency compared to traditional methods. The model performed better than conventional and deep learning baselines, having an accuracy of 98.9%, precision of 97.8%, recall of 97.9%, F1-score of 96.9%, ROC-AUC of 0.98, detection accuracy of 98.5%, false positive rate of 1.8%, and average latency of 42 ms, illustrating applicability to real-time use. The proposed DSVMTC model effectively improves real-time threat classification in VANET-based transportation systems. The hybrid architecture not only boosts detection accuracy but also ensures low-latency processing, making it suitable for real-world intelligent transportation networks. Future research will focus on optimizing the model for large-scale deployment and real-time adaptability.