摘要
针对现有K近邻非参数回归方法的局限,为了进一步提高算法的精度和速度,做出了两方面的改进:利用相关性理论选择状态向量和采用基于聚类分析的变K近邻搜索算法.用仿真实验验证了该方法的有效性,给出了仿真试验结果.实例研究结果表明,预测效果优于传统的非参数回归方法.
To address the shortage of the K-nearest Neighbor Nonparametric Regression, and to improve the accuracy and computing speed of the proposed algorithm, a method for improving the accuracy and computing speed is presented: choosing state vector based on self-association or association analysis and using an improved variable K searching method based on clustering analysis, simulation experiments are conducted to examine the validity of the method. The calculated results show that this method can produce more exact forecasting performance than the traditional one.
出处
《长沙交通学院学报》
2007年第4期39-43,共5页
Journal of Changsha Communications University
基金
国家自然科学基金资助项目(50478088)
关键词
交通工程
短时交通流量预测
非参数回归
相关性分析
聚类分析
traffic engineering
short-term traffic flow forecasting
nonparametric regression
association analysis
clustering analysis