Traffic flow prediction using support vector regression

Traffic flow prediction is a crucial measure in Intelligent Transportation System. It helps in efficiently handling the future vehicular load on the roads that will assist in managing traffic, reducing congestions and accident rates. Therefore, this study has been conducted on Jawaharlal Nehru University (JNU) located in New Delhi, India that covers 1019.38 acres of campus land. This paper considersthe previously in-depth studied real-time vehicular traffic of JNU which was manually monitored, collected, calculated and analyzed. It containsJanuary 2013 digitized data of the north gate of the campus which consisted of 31 days of 24 h each. The traffic-flow using Support Vector Regression is predicted, as it demonstrates better generalization ability and gives global minima for training samples.Further, the root-mean-squared errorand mean absolute errorwere computed as statistical measures to test the accuracy of the flow prediction for both incoming and outgoing traffic. © 2022, Bharati Vidyapeeth's Institute of Computer Applications and Management.