The identification of the most important predictors of the analyzed target variable strongly affects the accuracy of its interpretation and prediction, and many methods have been proposed in the literature aiming at variable selection. To assess the importance of each predictor, we propose the use of two algorithmic models to construct specific measures: Predictive Importance and Constructive Importance. We apply this procedure using classification and regression trees (CART) to investigate the effects of specific job satisfaction facets on overall job satisfaction, using a sample of workers of public and private nonprofit organizations in the Italian social service sector.
Mining the drivers of job satisfaction using algorithmic variable importance measures
CARPITA, Maurizio;ZUCCOLOTTO, Paola
2008-01-01
Abstract
The identification of the most important predictors of the analyzed target variable strongly affects the accuracy of its interpretation and prediction, and many methods have been proposed in the literature aiming at variable selection. To assess the importance of each predictor, we propose the use of two algorithmic models to construct specific measures: Predictive Importance and Constructive Importance. We apply this procedure using classification and regression trees (CART) to investigate the effects of specific job satisfaction facets on overall job satisfaction, using a sample of workers of public and private nonprofit organizations in the Italian social service sector.File | Dimensione | Formato | |
---|---|---|---|
CARPITA 2008 Mining the drivers of job satisfaction.pdf
gestori archivio
Tipologia:
Full Text
Licenza:
DRM non definito
Dimensione
1.33 MB
Formato
Adobe PDF
|
1.33 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.