Least squares regression is highly unreliable when a strong collinearity structure is present among the predictors. Among several proposals introduced in the literature, principal component regression is a straightforward method to overcome the problem, even if it introduces a slight bias in the parameter estimation. This paper presents a simulation study to evaluate the use of principal component regression in the context of quantile regression and, focusing on the variability of the estimates and the model’s prediction ability.
The Use of Principal Components in Quantile Regression: a Simulation Study / Davino, Cristina; Næs, Tormod; Romano, Rosaria; Vistocco, Domenico. - (2023), pp. 410-413. (Intervento presentato al convegno 14th Scientific Meeting of the Classification and Data Analysis Group tenutosi a Salerno nel 11-13 settembre 2023).
The Use of Principal Components in Quantile Regression: a Simulation Study
Cristina Davino;Rosaria Romano
;Domenico Vistocco
2023
Abstract
Least squares regression is highly unreliable when a strong collinearity structure is present among the predictors. Among several proposals introduced in the literature, principal component regression is a straightforward method to overcome the problem, even if it introduces a slight bias in the parameter estimation. This paper presents a simulation study to evaluate the use of principal component regression in the context of quantile regression and, focusing on the variability of the estimates and the model’s prediction ability.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
