TRAFFIC CONGESTION PREDICTION BASED ON A CASE BASED MODELING METHOD

In this paper, the authors present a traffic congestion predicting method for local and short term prediction that is based on a case- based modeling method. Current traffic conditions as well as features such as day of the week and hour are used in the method. Real or observed traffic data sets are first stored with their features in a database. The future traffic condition is then retrieved from the database based on the distance of the features. Experimental results indicate that this methodology has a high accuracy rate in predicting traffic congestion.