predict.BGR: Predict Method for Bivariate Gamma Regressions

In senhu/mvClaim: Multivariate general insurance claims modelling

Description

Obtains predictions from the fitted bivariate gamma regression models.

Usage

## S3 method for class 'BGR'
predict(object, newdata, ...)

predict_BGR_EE(object, newdata, ...)

predict_BGR_EI(object, newdata, ...)

predict_BGR_IE(object, newdata, ...)

Arguments

object	a fitted model object of class "BGR" or "MBGR"
newdata	a data frame that prediction is based upon
...	further arguments passed to or from other methods.

Value

a list with components (depends on model class):

fit	predictions.
alpha1.fit	predicted alpha1 values.
alpha2.fit	predicted alpha2 values.
alpha3.fit	predicted alpha3 values.
beta.fit	predicted beta values.
tau.fit	predicted mixing proportion values, if covariates enter the gating network.

Examples

mod1 <- BGR(modelName = "EE",
            y = c("y1","y2"), data = fullsim,
            f1 = ~ w1 + w2,
            f2 = ~ w2 + w3,
            f3 = ~ w1 + w2 + w3,
            f4 = ~ w1 + w2 + w3,
            verbose= FALSE)
fitted1 <- predict(mod1, newdata=fullsim)
plot(fitted1$fit)

mod2 <- BGR(modelName = "EI",
            y = c("y1","y2"), data = fullsim,
            f1     = ~ w1 + w2,
            f2     = ~ w2 + w3,
            f3     = ~ w1 + w2 + w3,
            verbose= FALSE)
fitted2 <- predict(mod2, newdata=fullsim)
mod3 <- BGR(modelName = "IE",
            y = c("y1","y2"), data = fullsim,
            f4     = ~ w1 + w2 + w3,
            verbose= FALSE)
fitted3 <- predict(mod3, newdata=fullsim)
plot(fitted2$fit)
plot(fitted3$fit)

senhu/mvClaim documentation built on Jan. 29, 2022, 3:18 p.m.