vecbinsolv: Solve systems of nonlinear equations based on a monotonic...

In stephenslab/EbayesThresh: Empirical Bayes Thresholding and Related Methods

Description

Solve a nonlinear equation or a vector of nonlinear equations based on an increasing function in a specified interval.

Usage

vecbinsolv(zf, fun, tlo, thi, nits = 30, ...)

Arguments

zf	the right hand side of the equation(s) to be solved
fun	an increasing function of a scalar argument, or a vector of such functions
tlo	lower limit of interval over which the solution is sought
thi	upper limit of interval over which the solution is sought
nits	number of binary subdivisions carried out
...	additional arguments to the function fun

Details

If fun is a scalar monotone function, the routine finds a vector t the same length as zf such that, component-wise, fun(t) = zf, where this is possible within the interval (tlo, thi). The relevant value returned is the nearer extreme to the solution if there is no solution in the specified range for any particular component of zf. The routine will also work if fun is a vector of monotone functions, allowing different functions to be considered for different components. The interval over which the search is conducted has to be the same for each component.

The accuracy of the solution is determined by the number of binary subdivisions; if nits=30 then the solution(s) will be accurate to about 9 orders of magnitude less than the length of the original interval (tlo, thi).

Author(s)

Bernard Silverman

References

See ebayesthresh and http://www.bernardsilverman.com

stephenslab/EbayesThresh documentation built on May 15, 2019, 4:28 p.m.