| Autor: |
SRIPRIYA, THODUR PARTHASARATHY, SRINIVASAN, MAMANDUR RANGASWAMY, SUBBIAH, MEENAKSHISUNDARAM |
| Předmět: |
|
| Zdroj: |
Chilean Journal of Statistics (ChJS); Apr2020, Vol. 11 Issue 1, p25-39, 15p |
| Abstrakt: |
Detecting outliers in two-way contingency tables is an important and interesting statistical problem. There is no clear objective procedure available in literature to handle outliers in categorical data unlike other data types. Therefore, this study envisages a two-step procedure, to first indicate and then to identify outliers in two-dimensional contingency tables. The approach deals with enhancing the summary measure to indicate the presence of possible outlying cells followed by residual approaches supplemented by boxplot in identifying the outliers. The fundamental definition of outlying cell as "markedly deviant" cell is clearly exploited in this two-step procedure. A simulation study has been carried out to examine the consistency of the proposed methods and later applied to a large collection of real datasets from various applications of social sciences. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
