On confidence intervals obtained from the fusion of samples

S. Fernández Alonso, J. Baz González, S. Diaz-Vazquez, S. Montes

Confidence intervals (CIs) are built to estimate an unknown parameter. Once a sample is observed, a CI becomes a specific interval of real values.

When we have isolated values, the aggregation of information from different sources has been widely studied. However, the study of interval aggregation is still in its infancy.

In this contribution, we study the fusion of sample data from a population. Specifically, we consider several independent samples of a variable and compare the individual confidence intervals for the mean with the confidence interval obtained from the pooled sample. We consider the latter one as an aggregation of the former ones and study general properties from the perspective of aggregating operators. In particular, we prove that it is an aggregation function for intervals in a formal sense if the appropriate order for intervals is considered.

Keywords: aggregation function, confidence interval, fusion sample, interval ordering

Data Analysis II
June 10, 2025  3:30 PM
Auditorio 2. Leandre Cristòfol

C. Gandia Tortosa, M. D. Molina Vila, M. J. Nueda Roldán

N. Amama Ben Hassun, J. Cortés Martínez, D. Fernández

C. Perez Lopez, A. RECIO ALCAIDE, J. A. Martín Segura

B. Pérez Sánchez, C. Perea, M. González Espinosa, J. J. López-Espín