EXPLORATION AND ANALYSIS OF TIME SERIES MODELS FOR INTELLIGENT TRAFFIC MANAGEMENT SYSTEM

Abstract

Traffic flow prediction is a research topic signified by several researchers in a league span of disciplines. Traffic flow prediction is an important aspect in Intelligent Transport Management System (ITMS). In this context, one of the most in-demand techniques of Machine Learning, especially Time series based techniques, helps in predicting traffic flow forecasting and increases the accuracy of the prediction model. In order to deliver extremely precise traffic forecasts, it is crucial that we put the prediction system into practice in the actual world. Our aim is to perform computations related to traffic on the traffic datasets and find out the accuracy for each model. For this purpose we are using three distinct time series models: Long Short Term Memory (LSTM), the Autoregressive Integrated Moving Average (ARIMA), and the Seasonal Autoregressive Integrated Moving Average (SARIMA). From the results obtained, it is concluded that the proposed model achieves highest prediction accuracy with the lowest root mean squared error.

Keywords

• itms
• lstm; arima
• seasonal arima
• rmse