computeBoundary: A function to compute the boundary of the decision region in...

In bnaras/sglr: Sequential Generalized Likelihood Ratio boundaries for pre-licensure vaccine trials

Description

This function computes the boundary of the decision region in a manner that can be employed in the field, as a table, for example. See section 4.2 of the reference below.

Usage

computeBoundary(b1, b0, p, glrTables = NULL, tol=1e-7)

Arguments

b1	The acceptance boundary value (corresponds to the boundary b_1 in the appendix of reference)
b0	The rejection boundary value (corresponds to the boundary b_0 in the appendix of reference)
p	The vector of probabilities, (p_0, p_1) with p_0 < p_1.
glrTables	A previously computed set of likelihood functions, to speed up computation for the same hypothesis testing problem. Otherwise, it is computed ab initio, resulting in a longer running time.
tol	A numerical tolerance, defaults to 1e-7

Details

This essentially computes the probabilities of hitting the boundaries using a recursion.

Value

upper	The upper boundary that indicates rejection of the null hypothesis
lower	The upper boundary that indicates acceptance of the null hypothesis
estimate	The estimated α and β values corresponding to the two boundaries

Author(s)

Balasubramanian Narasimhan

References

Please consult the website http://med.stanford.edu/biostatistics/vaccines.html for further developments. Particularly see the paper “Sequential Generalized Likelihood Ratio Tests for Vaccine Safety Evaluation” on the website that is due to appear in Statistics in Medicine.

See Also

See Also glrSearch

Examples

computeBoundary(b1=2.8, b0=3.3, p=c(.5, .75))

bnaras/sglr documentation built on May 12, 2019, 11:27 p.m.