globalweights_diff: Global Weights Based on Response Rate Differences
Description
Global Weights Based on Response Rate Differences
Usage
globalweights_diff(n, r, eps_global = 1, w = 1)
Value
A numeric value.
Arguments
n
The sample size per basket.
r
Vector of responses.
eps_global
A tuning parameter that determines the amount of borrowing.
A higher value leads to a smaller weight and therefore less borrowing
when the heterogeneity between the results in the baskets increases.
w
A fixed probability between 0 and 1. w is multiplied to
the weight.
Details
globalweights_diff calculates a weight based on the
heterogeneity of the response rates of all baskets that is multiplied
to the pairwise weights calculated with the function that is passed to
weight_fun. The weight is 1 when the number of responses is identical
in all baskets and 0 if the response rates are an equidistant sequence
from 0 to 1. If the maximum weight should be smaller than 1, w
can be set to a smaller value.