Missing data imputataion for traffic flow based on weighted local least squres

Data missing is the issue that is encounter in Intelligent Transportation System (ITS). The missing problem gives rise to many difficulties to further utilization. To overcome these difficulties, various methods have been proposed for the missing data imputation. A imputation method based on weighted local least squares for the missing data imputation of traffic flow is proposed in the paper. The real-time flow data recorded by SCOOT system in Beijing is applied to test the performance of the method and experiments results shown that the raised method has better imputing capabilities than BPCA imputation method.