rgnoisefilt

CRANrgnoisefiltMaRDI QIDQ5981618FDOQ5981618

Elimination of Noisy Samples in Regression Datasets using Noise Filters

Emilio Corchado, José A. Sáez, Juan Martin

Last update: 2 October 2023

Copyright license: GNU General Public License

Software version identifier: 1.1.1, 1.1.2

Description

Traditional noise filtering methods aim at removing noisy samples from a classification dataset. This package adapts classic and recent filtering techniques for use in regression problems, and it also incorporates methods specifically designed for regression data. In order to do this, it uses approaches proposed in the specialized literature, such as Martin et al. (2021) [<doi:10.1109/ACCESS.2021.3123151>] and Arnaiz-Gonzalez et al. (2016) [<doi:10.1016/j.eswa.2015.12.046>]. Thus, the goal of the implemented noise filters is to eliminate samples with noise in regression datasets.

Cites work

• On the Regressand Noise Problem: Model Robustness and Synergy With Regression-Adapted Noise Filters
• Instance selection for regression by discretization

This page was built for software: rgnoisefilt