Outlier Detection in Classification Analysis: An Overview | ||||
| The Egyptian Statistical Journal | ||||
| Article 5, Volume 37, Issue 2, December 1993, Page 222-237 | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/esju.1993.314842 | ||||
 View on SCiNiTO | ||||
| Author | ||||
| Ramses Fouad Sadek | ||||
| Abstract | ||||
| Some of the techniques used for detecting multivariate outliers can be used for detecting outliers in classification analysis. However, detecting outliers in classification analysis is more complicated than in any other analysis since the impact of an outlier on both group means and covariance matrices must be evaluated. In the current article, the goal is to give an overview of the multivariate outlier detection methods which can be employed in classification analysis. Outlier detection techniques include graphical and inferential methods. Graphical methods may not require any distributional assumptions regarding the data, while inferential methods assume a certain statistical model of the data. A brief summary of the popular graphical and inferential techniques used in multivariate outlier detection, which could be employed with classification analysis is presented. Comments and remarks on the outlier detection methods are also given. | ||||
| Keywords | ||||
| Classification; Discriminant Analysis; Mahalanobis Distance; Multivariate Analysis; Outliers | ||||
|
Statistics Article View: 34 |
||||
