rxTweedie: Tweedie Generalized Linear Models


Produces a dummy generalized linear model family object that can be used with rxGlm to fit Tweedie generalized linear regression models. This does NOT produce a full family object, but gives rxGlm enough information to call a C++ implementation that fits a Tweedie model. The excellent R package tweedie by Gordon Smyth does provide a full Tweedie family object, and that can also be used with rxGlm.


  rxTweedie(var.power = 0, link.power = 1 - var.power)



index of power variance function.


index of power link function. Setting link.power to 0 produces a log link function. Setting it to 1 is the identity link. The default is a canonical link equal to 1 - var.power


This provides a way to specify the var.power and the link.power arguments to the Tweedie distribution for rxGlm.


Microsoft Corporation Microsoft Technical Support


Peter K Dunn (2011). tweedie: Tweedie exponential family models. R package version 2.1.1.

See Also



 # In the claims data, the cost is set to NA if no claim was made
 # Convert NA to 0 for the cost, and read data into a data frame

 claimsXdf <- file.path(rxGetOption("sampleDataDir"),"claims.xdf")
 claims <- rxDataStep(inData = claimsXdf, 
            transforms = list(cost = ifelse(is.na(cost), 0, cost)))

 # Estimate using a Tweedie family                           
 claimsTweedie <- rxGlm(cost ~ age + car.age + type , 
     data=claims, family = rxTweedie(var.power = 1.5)) 

 # Re-estimate using a Tweedie family setting link.power to 0,
 # resulting in a log link function                          
 claimsTweedie <- rxGlm(cost ~ age + car.age + type , 
     data=claims, family = rxTweedie(var.power = 1.5, link.power = 0))