Bivariate Beta-binomial Regression Models
biv.betab fits dependent (logit) linear regression models to a
bivariate beta-binomial distribution.
biv.betab(freq, x = NULL, p, depend = TRUE, print.level = 0, typsize = abs(p), ndigit = 10, gradtol = 1e-05, stepmax = 10 * sqrt(p %*% p), steptol = 1e-05, iterlim = 100, fscale = 1)
freq |
A matrix containing four columns corresponding to 00, 01, 10, and 11 responses. |
x |
A matrix of explanatory variables, containing pairs of columns, one for each response, and the same number of rows as freq. |
p |
Initial parameter estimates: intercept, dependence (if depend is TRUE, and one for each pair of columns of x. |
depend |
If FALSE, the independence (logistic) model is fitted. |
print.level |
Arguments for nlm. |
typsize |
Arguments for nlm. |
ndigit |
Arguments for nlm. |
gradtol |
Arguments for nlm. |
stepmax |
Arguments for nlm. |
steptol |
Arguments for nlm. |
iterlim |
Arguments for nlm. |
fscale |
Arguments for nlm. |
A list of class bivbetab is returned.
J.K. Lindsey
y <- matrix( c( 2, 1, 1,13, 4, 1, 3, 5, 3, 3, 1, 4, 15, 8, 1, 6),ncol=4,byrow=TRUE) first <- c(0,0,1,1) second <- c(0,1,0,1) self <- cbind(first,second) other <- cbind(second,first) biv.betab(y,cbind(self,other),p=c(-1,2,1,1)) # independence biv.betab(y,cbind(self,other),p=c(-1,1,1),dep=FALSE)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.