The function BS() defines The Birnbaum-Saunders,
a two parameter distribution, for a gamlss.family object
to be used in GAMLSS fitting
using the function gamlss().
Value
Returns a gamlss.family object which can be used to fit a
BS distribution in the gamlss() function.
Details
The Birnbaum-Saunders with parameters mu and sigma
has density given by
\(f(x) = \frac{x^{-3/2}(x+\mu)}{2\sigma\sqrt{2\pi\mu}} \exp\left(\frac{-1}{2\sigma^2}(\frac{x}{\mu}+\frac{\mu}{x}-2)\right)\)
for \(x>0\), \(\mu>0\) and \(\sigma>0\). In this parameterization \(\mu\) is the median of \(X\), \(E(X)=\mu(1+\sigma^2/2)\) and \(Var(X)=(\mu\sigma)^2(1+5\sigma^2/4)\). The functions proposed here corresponds to the functions created by Roquim et al. (2021) with minor modifications to obtain correct log-likelihoods and random samples.
References
Roquim, F. V., Ramires, T. G., Nakamura, L. R., Righetto, A. J., Lima, R. R., & Gomes, R. A. (2021). Building flexible regression models: including the Birnbaum-Saunders distribution in the gamlss package. Semina: Ciências Exatas e Tecnológicas, 42(2), 163-168.
Examples
# Example 1
# Generating some random values with
# known mu and sigma
y <- rBS(n=100, mu=0.75, sigma=1.3)
# Fitting the model
require(gamlss)
mod1 <- gamlss(y~1, sigma.fo=~1, family=BS)
#> GAMLSS-RS iteration 1: Global Deviance = 237.3994
#> GAMLSS-RS iteration 2: Global Deviance = 237.3972
#> GAMLSS-RS iteration 3: Global Deviance = 237.3972
# Extracting the fitted values for mu and sigma
# using the inverse link function
exp(coef(mod1, what="mu"))
#> (Intercept)
#> 0.7713414
exp(coef(mod1, what="sigma"))
#> (Intercept)
#> 1.243161
# Example 2
# Generating random values for a regression model
# A function to simulate a data set with Y ~ BS
gendat <- function(n) {
x1 <- runif(n)
x2 <- runif(n)
mu <- exp(1.45 - 3 * x1)
sigma <- exp(2 - 1.5 * x2)
y <- rBS(n=n, mu=mu, sigma=sigma)
data.frame(y=y, x1=x1, x2=x2)
}
set.seed(123)
dat <- gendat(n=300)
mod2 <- gamlss(y~x1, sigma.fo=~x2,
family=BS, data=dat)
#> GAMLSS-RS iteration 1: Global Deviance = 6624.679
#> GAMLSS-RS iteration 2: Global Deviance = 6593.565
#> GAMLSS-RS iteration 3: Global Deviance = 6556.349
#> GAMLSS-RS iteration 4: Global Deviance = 6506.77
#> GAMLSS-RS iteration 5: Global Deviance = 6432.284
#> GAMLSS-RS iteration 6: Global Deviance = 6279.265
#> GAMLSS-RS iteration 7: Global Deviance = 1685.61
#> GAMLSS-RS iteration 8: Global Deviance = 1208.005
#> GAMLSS-RS iteration 9: Global Deviance = 1202.153
#> GAMLSS-RS iteration 10: Global Deviance = 1202.152
summary(mod2)
#> Warning: summary: vcov has failed, option qr is used instead
#> ******************************************************************
#> Family: c("BS", "Birnbaum-Saunders")
#>
#> Call: gamlss(formula = y ~ x1, sigma.formula = ~x2, family = BS, data = dat)
#>
#>
#> Fitting method: RS()
#>
#> ------------------------------------------------------------------
#> Mu link function: log
#> Mu Coefficients:
#> Estimate Std. Error t value Pr(>|t|)
#> (Intercept) 1.476 0.236 6.256 1.37e-09 ***
#> x1 -3.274 0.376 -8.707 < 2e-16 ***
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> ------------------------------------------------------------------
#> Sigma link function: log
#> Sigma Coefficients:
#> Estimate Std. Error t value Pr(>|t|)
#> (Intercept) 2.01922 0.08094 24.95 <2e-16 ***
#> x2 -1.52035 0.14010 -10.85 <2e-16 ***
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> ------------------------------------------------------------------
#> No. of observations in the fit: 300
#> Degrees of Freedom for the fit: 4
#> Residual Deg. of Freedom: 296
#> at cycle: 10
#>
#> Global Deviance: 1202.153
#> AIC: 1210.153
#> SBC: 1224.968
#> ******************************************************************
# Example 3