Study of outliers under variance-inflation model in experimental designs
2003
Bhar, L. | Gupta, V.K. (Indian Aricultural Statistics Research Institute, New Delhi (India))
For detecting outliers in linear regression model,Cook developed[2] a statistic under 'mean-shift' model. This statistic is appropriately developed under 'variance-inflation' model for detecting outliers in yhe general linear model for experimental designs. The simplified form of this stastic is obtained for both one-way and two-way elimination of heterogeneity settings. the relationship between Cook-stastic under 'mean-shift' model and Cook-stastic under 'variance-inflation' model for a single outlier is established and it has been shown that an observation which is an outlier under'mean shift' model may not be so under "variance-inflation' model. Throgh an example it has shown that heterscedasticity in the data may not infuence the estimation of parameters.
Mostrar más [+] Menos [-]Palabras clave de AGROVOC
Información bibliográfica
Este registro bibliográfico ha sido proporcionado por Indian Council of Agricultural Research