Use the empirical likelihood ratio and Wilks theorem to test if the
regression coefficient is equal to beta0
,
by the case weighted estimation method.
The log empirical likelihood been maximized is $$ \sum_{d=1} \log \Delta F(y_i) + \sum_{d=0} \log [1-F(y_i)].$$
WRegTest(x, y, delta, beta0, psifun=function(t){t})
A list with the following components:
the -2 log likelihood ratio; have approximate chisq distribution under \(H_0\).
the p-value using the chi-square approximation.
a matrix of size N by q. Random design matrix.
a vector of length N, containing the censored responses.
a vector (length N) of either 1's or 0's. delta=1 means y is uncensored; delta=0 means y is right censored.
a vector of length q. The value of the regression coefficient to be tested in the linear model
.
the estimating function. The definition of it determines the type of estimator under testing.
Mai Zhou.
The above likelihood should be understood as the likelihood of the
censored responses y
and delta
.
This version can handle the model where beta is a vector (of length q).
The estimation equations used when maximize the empirical likelihood is $$ 0 = \sum \delta_i \Delta F(Y_i) X_i \psi( Y_i - X_i \beta0 ) $$ which was described in detail in the reference below.
For median regression (Least Absolute Deviation) estimator, you should
define the
psifun
as \(+1, -1\) or \(0\) when \(t\) is \(>0, <0 \)
or \( =0\).
For ordinary least squares estimator, psifun
should be the identity function psifun <- function(t)t.
Zhou, M.; Kim, M. and Bathke, A. (2012). Empirical likelihood analysis of the case weighted estimator in heteroscastic AFT model. Statistica Sinica, 22, 295-316.
xx <- c(28,-44,29,30,26,27,22,23,33,16,24,29,24,40,21,31,34,-2,25,19)
Run the code above in your browser using DataLab