Abstract

Abstract:- The traffic flow prediction has wide application in the city transportation and area management. In big cities, it is very difficult to manage traffic. But the prediction under consideration of some physical conditions of environment and weather is found more effective. In this study, we designed a traffic flow prediction system model to predict the traffic data with a time interval of 1 hour to 24 hours. The prediction algorithms have been used for research in the past, but there are not many platforms found on which traffic flow prediction has easy access to public users. The system is designed to overcome the problems associated with the historical and time series. Past Traffic data set was collected from an open-source and cleaned as per requirements. By using Machine learning algorithms, a system is designed, which takes the data from the roads using Vehicle detection sensors and stores into the database for future predictions. We also integrated the API of weather services to get weather data. This Traffic flow prediction Model is designed to use the existing two most popular machine learning prediction algorithms that Support Vector Machine (SVM).After experiments, results were compared with the actual data to check the accuracy of the algorithms. Support Vector Machine (SVM) helps to predict in short term prediction scenarios. But a shorter time interval gives more accurate results.Key Words: SVM, Machine Learning, pre-processing, Classification, deep learning