Recordtest: an R package to analyze non-stationarity in the extremes based on record-breaking events

Castillo-Mateo, Jorge (Universidad de Zaragoza) ; Cebrián, Ana C. (Universidad de Zaragoza) ; Asín, Jesús (Universidad de Zaragoza)

Resumen: The study of non-stationary behavior in the extremes is important to analyze data in environmental sciences, climate, finance, or sports. As an alternative to the classical extreme value theory, this analysis can be based on the study of record-breaking events. The R package RecordTest provides a useful framework for non-parametric analysis of non-stationary behavior in the extremes, based on the analysis of records. The underlying idea of all the non-parametric tools implemented in the package is to use the distribution of the record occurrence under series of independent and identically distributed continuous random variables, to analyze if the observed records are compatible with that behavior. Two families of tests are implemented. The first only requires the record times of the series, while the second includes more powerful tests that join the information from different types of records: upper and lower records in the forward and backward series. The package also offers functions that cover all the steps in this type of analysis such as data preparation, identification of the records, exploratory analysis, and complementary graphical tools. The applicability of the package is illustrated with the analysis of the effect of global warming on the extremes of the daily maximum temperature series in Zaragoza, Spain.

Idioma: Inglés

DOI: 10.18637/jss.v106.i05

Año: 2023

Publicado en: Journal of Statistical Software 106, 5 (2023), 1-28
ISSN: 1548-7660

Factor impacto JCR: 5.4 (2023)

Categ. JCR: STATISTICS & PROBABILITY rank: 4 / 168 = 0.024 (2023) - Q1 - T1
Categ. JCR: COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS rank: 30 / 170 = 0.176 (2023) - Q1 - T1

Factor impacto CITESCORE: 10.7 - Software (Q1) - Statistics, Probability and Uncertainty (Q1) - Statistics and Probability (Q1)

Factor impacto SCIMAGO: 2.709 - Software (Q1) - Statistics, Probability and Uncertainty (Q1) - Statistics and Probability (Q1)

Financiación: info:eu-repo/grantAgreement/ES/DGA/E46-20R
Financiación: info:eu-repo/grantAgreement/ES/MICINN/PID2020-116873GB-I00

Tipo y forma: Article (Published version)

Área (Departamento): Área Estadís. Investig. Opera. (Dpto. Métodos Estadísticos)