Robust distance-based imputation techniques.

M. Álvarez Martín, E. Boj del Val, A. Grané Chávez

In this work we explore distance-based methodologies for data imputation in complex data sets of mixed-type data. Our proposal is based on the use of robust distances, calculated with the dbrobust R package, which allows combining numerical and categorical variables while reducing the influence of outliers. In particular, we analyze several real data sets with varying percentages of missing data and evaluate the efficiency and computing time of our proposal and some competitors. The results show that robust methods offer efficient missing value imputation.

Palabras clave: data imputation, dbrobust, mixed-type data, robust distances

GT AMyC I: Advances in Distance-Based Methods
4 de septiembre de 2026  09:00
Aula 28

M. Sánchez Ronco, S. Olivera, A. Grané

F. Scielzo Ortiz, A. Grané, M. Diaz Gorfinkiel

P. F. Delicado Useros, A. Casanova Lloveras, C. Pachón García

S. Díez Martínez, A. Arribas Gil, A. Grané Chávez