newton.theta: Initial theta value by Newton method

In sporm: Semiparametric Proportional Odds Rate Model

Description

Optional initial θ value for mrle.sporm based on uniform (0, 1) baseline distrbution and calculated by Newton method.

Usage

newton.theta(y, theta0 = 1, maxit = 100, eps = 1e-10)

Arguments

y	The y-sample in proportional odds rate model with uniform (0,1) baseline distrbution. If baseline F is not uniform (0,1) or unknown, use y^*_i=F_n(y_j), j=1,...,n, where F_n is the empirical cdf of x1,...,xm.
theta0	an initial value of θ
maxit	The maximum number of Newton iterations.
eps	Convergence tolerance used in the Newton iteration

Details

See the reference below.

Value

Returns the proportionality parameter θ of the parametric proportional odds rate model with Uniform(0,1) baseline by Newtom method.

Author(s)

Zhong Guan <[e-mail]>

References

Zhong Guan and Cheng Peng (2011), "A rank-based empirical likelihood approach to two-sample proportional odds model and its goodness-of-fit", Journal of Nonparametric Statistics, to appear.

See Also

mrle.sporm.

Examples

theta<-2
u<-runif(30)
y<-u/(theta-(theta-1)*u)
newton.theta(y)

sporm documentation built on Jan. 27, 2021, 5:08 p.m.