A Robust Method for Multiple Outliers Detection in Multi-Parametric Models
Article
Article Title | A Robust Method for Multiple Outliers Detection in Multi-Parametric Models |
---|---|
ERA Journal ID | 4653 |
Article Category | Article |
Authors | Chong, Albert K. |
Journal Title | Photogrammetric Engineering and Remote Sensing |
Journal Citation | 53 (6), pp. 617-620 |
Number of Pages | 4 |
Year | Jun 1987 |
Place of Publication | United States |
ISSN | 0099-1112 |
2374-8079 | |
Abstract | An analytical method consisting of concepts from the Danish robustified least squares, Pope's TAU statistic, and stepwise analysis is discussed. Simulated data were used to test the efficiency and reliability of the method. The collinearity condition in photogrammetry was the basic mathematical model. A flow diagram of the method and a summary of test results of test samples are included to provide more information on the method. The method can detect up to (11 - (11 +1/)/2 outliers in small samples (where 11 is the sample size and /I is the number of unknowns in the mathematical model). In large samples, more than II - (/1 +u)/2 outliers may be detectable. Geometric checkpoints were also used to reduce the effect of poor geometry of selected observations (especially in photogrammetry). |
Public Notes | There are no files associated with this item. |
Byline Affiliations | No affiliation |
https://research.usq.edu.au/item/w26x3/a-robust-method-for-multiple-outliers-detection-in-multi-parametric-models
30
total views0
total downloads2
views this month0
downloads this month