On zero-inflated permutation testing and some related problems
| Autor: | Fortunato Pesarin, Livio Finos |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Statistics and Probability
Multivariate statistics Interpretation (logic) Theoretical computer science Multivariate tests Permutation tests Zero-inflated Computer science Subject (documents) Test (assessment) Zero (linguistics) Permutation Covariate Permutation testing Statistics::Methodology Statistics Probability and Uncertainty |
| Zdroj: | Statistical Papers. 61:2157-2174 |
| ISSN: | 1613-9798 0932-5026 |
| DOI: | 10.1007/s00362-018-1025-x |
| Popis: | The inferences on zero-inflated data are difficult to deal and the problem motivated a relevant part of the research since the earlier times of the statistical science. The case of multivariate zero-inflated data is still subject of active debates. In this contribution we primarily deal with a permutation-based test for comparisons of two groups with multivariate zero-inflated data. By the use of a leading example, we formulate different questions and translate them on different inferential hypotheses. A permutation-based solution is proposed for each of them and their interpretation is discussed. Finally, we extend the method to the general case of—possibly many—continuous predictors and the presence of covariates (nuisance). The data and the R code are implemented in the library flip on CRAN repository and on the web-appendinx of this paper. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink