In this paper, a local regression imputation method (LRL) is proposed. For each nonrespondent, the regression model parameters are estimated by using the data concerning a set of its respondent neighbours. The aim of this procedure, is to make the estimates obtained by regression imputation methods robust against violations of the basic assumption that could be ignored during the esplorative phase of the sampling data.
Imputazione per regressione locale in presenza di non risposta non ignorabile / E. Rocco. - STAMPA. - (2003), pp. 1-4. (Intervento presentato al convegno Convegno intermedio della Società Italiana di Statistica - Analisi Statistica Multivariata per le Scienze Economiche-Sociali, le Svienze Naturali e la Tecnologia tenutosi a Napoli nel 9-11 Giugno 2003).
Imputazione per regressione locale in presenza di non risposta non ignorabile
ROCCO, EMILIA
2003
Abstract
In this paper, a local regression imputation method (LRL) is proposed. For each nonrespondent, the regression model parameters are estimated by using the data concerning a set of its respondent neighbours. The aim of this procedure, is to make the estimates obtained by regression imputation methods robust against violations of the basic assumption that could be ignored during the esplorative phase of the sampling data.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.