“…Likewise, model complexity and high dependency to a lot of data are the most significant disadvantages of them. Neural networks such as MLP, RBF and TDNN (Bin, 2006, Nagare, 2012, K Nearest Neighborhood (Chang, 2012) and SVM models (Castro-Neto, 2009) are among the most common nonparametric prediction methods that have been used more often for traffic flow prediction.…”