Short-term Traffic Flow Forecasting based on Two-tier K-nearest Neighbor Algorithm

Abstract

The K-nearest Neighbor algorithm does not require a priori knowledge and its forecasting results are better than those of the linear model algorithm. However, its computing speed is low and its parameter adjustment method is not flexible enough. Based on the traditional K-nearest neighbor algorithm, this paper proposes a two-tier K-nearest neighbor algorithm. Combined with the actual traffic flow, it calibrates the algorithm parameter to improve the calculation speed and the accuracy of the algorithm.