Chinese Space Science and Technology

• •    

An Improved Algorithm for the Rejection of Outlier Data

Tong Li Zhou Haiyin(National University of Defense Technology,Changsha 410073)   

  • Online:2001-08-25
  • Supported by:
    国家自然科学基金 (69872 0 3 9)资助;; 教育部高等院校骨干教师基金资助

Abstract: The recognition of massive outlier data is a problem with a large number of operations in data processing.As for a large-scaled linear regression model,a fast algorithm to pick out outliers is presented based on the one-by-one method.This algorithm can not only improve the recognition of outliers,especially for massive outliers,but also reduce the computational operations remarkably.