Performance Evaluation of Machine Learning Algorithms applied in SD-VANET for Efficient Transmission of Multimedia Information

For the advancement of technologies in vehicular industry, an intelligent Software Defined Vehicular Ad-hoc Network (SDVANET) which decoupled its data and control plane providing a golden opportunity to the researchers and academicians working in the related field. The limitation associated with SDVANET lies in the fact that it has problem in transferring reliable multimedia information among different users. To address this limitation, researchers of the domain tried to provide more intelligence to SDVANET by deploying Machine Learning (ML) algorithms to this network. This deployment of ML algorithms provided a safe, reliable, secure, optimal methodology for transferring multimedia information among SDVANET networks. In this work, researchers have used two datasets: small dataset and big dataset according to number of iterations used for collection. Total six supervised ML algorithms: Random Forest (RF), Decision-Tree (DT), Logistic Regression (LR), Naive-Bayes (NB), Support Vector Machine (SVM), K-Nearest Neigbour (kNN) are chosen to participate in the training and testing process. For result analysis, two simulation parameters: Receiver operating characteristic-Area under the ROC curve (ROC-AUC) and Confusion matrix are utilized. Numerical Results: Classification Accuracy (CA) (88.3%), F1-Measure (89.3%), precision (94.3%), recall (93.3%), AUC (92.3%) indicate that ML algorithms used for small dataset (LR, SVM and NB), NB is outperforming and ones used for big dataset (DT, RF and kNN), DT gives best performance based on numerical results: CA (95.4%), F1-Measure (93.4%), precision (92.3%), recall (98.2%), AUC (94.9%). So, NB and DT are the best ML algorithms for transferring the secure multimedia information, avoiding accidents among SDVANET networks.