The problem of outliers have been a important of regression analysis, because it will distort the meaning of whole data of its analysis and reduce the accuracy and validity for adequate models.
This paper of concerned with the identification of multi...
The problem of outliers have been a important of regression analysis, because it will distort the meaning of whole data of its analysis and reduce the accuracy and validity for adequate models.
This paper of concerned with the identification of multiple outliers. Three procedures for identifying outliers are illustrated and compared ; procedure based on the Hadi & Simonoff, procedure based on the robust estimator with LMS(least median squares) and Huber's M estimator.