Generalized Leverage Values

Compute the generalized leverages values for fitted models.

Usage

gleverage(model, ...)

Arguments

model

a model object.

...

further arguments passed to methods.

Value

gleverage is a new generic for computing generalized leverage values as suggested by Wei, Hu, and Fung (1998). Currently, there is only a method for betareg models, implementing the formulas from Rocha and Simas (2011) which are consistent with the formulas from Ferrari and Cribari-Neto (2004) for the fixed dispersion case.

Currently, the vector of generalized leverages requires computations and storage of order \(n \times n\).

References

Ferrari SLP, Cribari-Neto F (2004). Beta Regression for Modeling Rates and Proportions. Journal of Applied Statistics, 31(7), 799–815.

Rocha AV, Simas AB (2011). Influence Diagnostics in a General Class of Beta Regression Models. Test, 20(1), 95–119. doi:10.1007/s11749-010-0189-z

Wei BC, Hu, YQ, Fung WK (1998). Generalized Leverage and Its Applications. Scandinavian Journal of Statistics, 25, 25–37.

See also

betareg

Examples

options(digits = 4)
data("GasolineYield", package = "betareg")
gy <- betareg(yield ~ batch + temp, data = GasolineYield)
gleverage(gy)
#>      1      2      3      4      5      6      7      8      9     10     11 
#> 0.2167 0.2517 0.3254 0.4542 0.2239 0.3201 0.5271 0.2819 0.3011 0.5066 0.1970 
#>     12     13     14     15     16     17     18     19     20     21     22 
#> 0.2146 0.3054 0.4397 0.2909 0.3514 0.4049 0.2448 0.3570 0.4840 0.2154 0.1835 
#>     23     24     25     26     27     28     29     30     31     32 
#> 0.2899 0.4701 0.2910 0.2982 0.5449 0.3677 0.6603 0.3181 0.2557 0.4569