Tuning of Machine Learning Model to Strengthen the Digital Ecosystem of Vehicular Ad-hoc Networks

The paper presents an attempt to improve cybersecurity by optimizing machine learning methods. In the process, authors present a dynamic method that enhances these algorithms, leading to a better accuracy in threat identification. This improves the security of digital ecosystems and provides information on how cybersecurity techniques are changing. The authors offer an in-depth analysis of security risks in vehicular ad hoc networks (VANETs) by dividing data into "attack," "moderate attack," and "no attack" segments using Mean Square Error (MSE). This method determines the impact and seriousness of the threats. Our work provides detailed insights for proactive security in VANETs by using MSE as a labelling technique. To enhance threat assessment, a rating system for vehicles is also introduced. This system assesses vehicle behavior and performance within VANETs using the Levenberg algorithm. By examining service measurements, a rank is assigned to every vehicle, signifying its influence on cybersecurity and network performance. The effectiveness of threat detection is examined using performance metrics demonstrating nearly 4% enhancement in threat detection accuracy. This combined strategy raises individual vehicle contributions to QoS in VANETs, boosts threat detection, and offers insights into the health of the network.