论文部分内容阅读
数据丢失是动态交通数据采集中的常见现象,丢失数据的补齐质量对交通信息的准确性具有重要影响。本文根据交通数据随时间变化的特点,利用交通检测器采集到的动态数据,对训练数据进行了4种插值方法的插值效果比较,进行了少量数据丢失和大量数据丢失的补齐方法研究,建立了拟合多项式函数,根据保形拟合确定了合理的补齐数值,补齐数据和原始数据的平均相对误差5.0%,数据补齐精度高。
Data loss is a common phenomenon in dynamic traffic data collection. The quality of missing data has an important influence on the accuracy of traffic information. According to the characteristics of traffic data changing with time, this paper uses the dynamic data collected by traffic detector to compare the interpolation results of training data with four kinds of interpolation methods, and studies the method of adding a small amount of data loss and large amount of data loss to establish The fitting polynomial function, according to conformal fitting to determine a reasonable padding value, padding data and the average relative error of the original data of 5.0%, the data padding accuracy.