The Prediction of Traffic Flow with Regression Analysis

Abstract

Road traffic management is necessary for smart cities. Lately, Intelligent Transport System (ITS) has become an important area of research to solve different road traffic related issues for making smart decision. It links people, roads and vehicles together using communication technologies to increase safety and mobility. Moreover, accurate prediction of road traffic is also important to manage traffic congestion. In this thesis work, we give a brief survey of recent researches on road traffic prediction and propose an innovative approach to estimate road traffic flow using regression analysis for the roads of Porto city (capital of Portugal). The 22 inductive loop traffic detectors (loop sensors) were used to collect traffic data around VCI (Via de Cintura Interna) motorway. Initially, we have applied data preprocessing and feature selection techniques on the traffic data and then applied several regression models (Linear Regression, SMO Regression, Multilayer Perceptron, M5P Regression Tree and Random Forest) to predict future traffic flow. Finally, we apply regression models including an ensemble model proposed in the study to predict road traffic flow in long-term based on historical traffic data and compare their performance. The experiment results show that prediction generated by M5P based regression tree tend to get the closest to actual traffic flow for different roads of Porto city, Portugal.