TRAFFIC FLOW PREDICTION USING MACHINE LEARNING

Abstract

As a result of the ongoing increase in the number of vehicles, there are increasingly severe traffic bottlenecks. The amount of time and money that users of transportation spend traveling will immediately change as a result. To some extent, this issue can be mitigated by projecting future traffic patterns. Three machine learning algorithms—Linear Regression, Decision Tree, and Support Vector Machine—are used in this work to predict the traffic flow after the data has been preprocessed using Selenium, OSS, and Message Queue. Next, we examine real-time Beijing traffic data as it shows up on the Baidu map. According to this study, Random Forest is the most accurate of all three techniques, with an accuracy of 0.719. Second are logistic regression and SVM. Keywords - Traffic FlowPrediction; Decision Tree; Random Forest; SVM; Baysian Ridge; CNN; LSTM