论文部分内容阅读
分析了原有的短时交通流预测的K近邻算法,用模式距离搜索方法代替原有的欧氏距离搜索方法,引入多元统计回归模型,建立了一种改进的短时交通流预测的K近邻算法,并以北京市某路段进行实例验证。试验结果表明:当K取23时,利用改进的K近邻算法,预测结果的均方误差、平均相对误差、平均绝对误差分别为31.43%、4.17%、0.27%;利用原有的K近邻算法,预测结果的均方误差、平均相对误差、平均绝对误差分别为33.33%、4.40%、0.28%;利用历史平均模型,预测结果的均方误差、平均相对误差、平均绝对误差分别为46.20%、11.40%、0.48%。可见,改进的K近邻算法的预测精度明显高于其他2种方法,在提高搜索效率的同时准确地刻画了交通流的真实情况。
The K-nearest neighbor algorithm for short-term traffic flow prediction was analyzed. The original Euclidean distance search method was replaced by the model distance search method, and multivariate regression model was introduced to establish an improved K-nearest neighbor Algorithm, and a section of Beijing to verify the examples. The experimental results show that the mean square error, average relative error and average absolute error of the predicted results are 31.43%, 4.17% and 0.27%, respectively, using the improved K nearest neighbor algorithm when K is taken as 23; using the original K nearest neighbor algorithm, The mean square error, average relative error and average absolute error of prediction results were 33.33%, 4.40% and 0.28% respectively. Using the historical average model, the mean square error, average relative error and average absolute error of the prediction results were 46.20% and 11.40 respectively %, 0.48%. It can be seen that the prediction accuracy of the improved K nearest neighbor algorithm is obviously higher than the other two methods, which accurately depicts the real situation of traffic flow while improving the search efficiency.