Traffic Flow Prediction: An Approach for Traffic Management

Transportation has always been a boon to humankind, and with the advent of motorized vehicles, humans can travel quicker and faster, but with the new solution arose new problems, these problems are usually caused by the bottlenecking of roads. These roads are not able to accommodate a huge number of vehicles; this, in turn, causes the whole transportation system sometimes to collapse. This issue is tried to be avoided, using a method called ‘Time-Series Forecasting’, using historic data and two statistical models namely, ARIMA and regression. These algorithms will be trained according to the data provided. The trained model could now be used, to predict the traffic conditions of a spot at a particular time. Comparing with ARIMA and regression model, ARIMA always becomes cost-effective and sometimes unable to set up in terminal stations. In this paper, the dataset has been restructured in such a way that the linear regression model itself be able to reach satisfactory performance.