getf0: f0 optimization routine

In gldrm: Generalized Linear Density Ratio Models

Description

f0 optimization routine

Usage

getf0(y, spt, ySptIndex, sptFreq, sampprobs, mu, mu0, f0Start, thStart,
  thetaControl = theta.control(), f0Control = f0.control(), trace = FALSE)

Arguments

y	Vector of response values.
spt	Vector of unique observed support points in the response.
ySptIndex	Index of each y value within spt.
sptFreq	Vector containing frequency of each spt value.
sampprobs	Optional matrix of sampling probabilities.
mu	Fitted mean for each observation. Only used if sampprobs=NULL.
mu0	Mean constraing for f0.
f0Start	Starting f0 values. (Typically the estimate from the previous iteration.)
thStart	Starting theta values. Needs to be a list of values matching the output of the getTheta function.
thetaControl	A "thetaControl" object returned from the theta.control function.
f0Control	An "f0Control" object returned from the f0.control function. trace Logical. If TRUE, then progress is printed to terminal at each iteration.

Value

A list containing the following:

• f0 Updated values.

• llik Updated log-likelihood.

• th Updated list returned from the getTheta function.

• conv Convergence indicator.

• iter Number of iterations until convergence.

• nhalf The number of half steps taken on the last iteration if the initial BFGS update did not improve the log-likelihood.

• score.log Score function with respect to log(f0) at convergence.

• info.log Information matrix with respect to log(f0) at convergence.

gldrm documentation built on May 2, 2019, 12:59 p.m.