A MULTI STREAM FEATURE FUSION APPROACH FOR TRAFFIC PREDICTION

Abstract

Traffic prediction has become an essential component of Intelligent Transportation Systems (ITS) due to the rapid increase in urbanization, vehicle population, and road congestion. Accurate traffic flow forecasting helps transportation authorities, commuters, and smart city systems make efficient decisions related to route planning, congestion management, fuel optimization, and accident prevention. However, traffic conditions are highly dynamic and nonlinear because they are influenced by various factors such as weather conditions, public events, road construction activities, holidays, and driving behavior. Traditional statistical and machine learning approaches often fail to capture complex spatial and temporal dependencies present in traffic data, leading to reduced prediction accuracy