Self-Organized Graph Neural Network with Adaptive Boosting Support Vector Regression Prediction Model for Smart Traffic Management System

Advancement in technology has led path for automated road traffic management software in order to minimize the road traffic. The prediction of road traffic helps to maintain the harmony in the society. The existing model failed to predict appropriate road traffic due to lack of data preprocess and feature extraction ability that affected the smart traffic management system. To Self-Organized Graph Neural Network with Adaptive Boosting Support Vector Regression (SOGNN-ABSVR) is proposed for the smart traffic management. The z-score normalization technique was used to normalize the ranges of the data that increased the model training. The Self-Organized Graph Neural Network (SOGNN) extracted the spatial and temporal features from the road network data. The Adaptive Boosting Support Vector Regression (ABSVR) classifier was used to predict the traffic network. The proposed SOGNN-ABSVR method performance was evaluated in terms of Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) performance parameters and has obtained greater results having RMSE of 4.97 for NYC dataset and 2.14 for PeMS-Bay datasets respectively compared to the existing Multi-Component Attention Graph Convolutional Network (MCAGCN) method.