The invention discloses a traffic data matrix filling method based on Hessian regular space-time low-rank constraint, and the method comprises the steps: obtaining incomplete traffic data, and building a traffic data matrix; establishing a low-rank matrix completion model based on factor matrix decomposition; optimizing the low-rank matrix completion model based on factor matrix decomposition by analyzing time sequence change characteristics and Hessian regular space similarity characteristics of traffic data; and performing low-rank matrix completion with space-time constraint, and recovering original traffic data. According to the method, the low-rank matrix completion method based on factor matrix decomposition is applied to the field of traffic data recovery, the spatial-temporal correlation and low-rank characteristics of the traffic data are fully mined, the precision of recovering the complete traffic data is improved, and particularly, the method has a good application effect on traffic data recovery in different missing modes.
本发明公开基于Hessian正则时空低秩约束的交通数据矩阵填充方法,获取不完整的交通数据,建立交通数据矩阵;建立基于因子矩阵分解的低秩矩阵补全模型;通过分析交通数据的时序变化特性和Hessian正则空间相似特性,对所述基于因子矩阵分解的低秩矩阵补全模型进行优化;进行加入时空约束的低秩矩阵补全,恢复出原始交通数据。本发明将一种基于因子矩阵分解的低秩矩阵补全方法应用于交通数据恢复领域,并充分挖掘交通数据的时空相关性和低秩特性,提高了恢复完整交通数据的精度,尤其对不同缺失模式下的交通数据修复具有很好的应用效果。
Traffic data matrix filling method based on Hessian regular space-time low-rank constraint
基于Hessian正则时空低秩约束的交通数据矩阵填充方法
2022-09-20
Patent
Electronic Resource
Chinese
A Rank Two Hessian Matrix Update for Sequential Quadratic Approximation
British Library Conference Proceedings | 1995
|Traffic speed data filling method based on space-time regularization tensor decomposition
European Patent Office | 2024
|