NEWS

gte 1.2-3 (2022-04-16)

Bug fix : gte() now deals properly with datasets containing no interval censored observations.

gte() now returns 'intmap', the intervals of the potential steps in the conditional survival function, called innermost interval, in which the GTE is indeterminate. They are computed with an internal function derived from function Aintmap() of the 'interval' package.
gte() also returns 'surv.summary', the estimate of the conditional survival function only for the innermost intervals.
A print and a plot method have been added for objects of class 'gte'.
In gte(), the initial death probability are now null for intervals not among the innermost intervals.
The values in the gte() output vector 'time', have been shifted down one position to be correctly associated with the survival values in the output matrix 'surv'.
Documentation update and improvement.
Tests in inst/tests/ have been updated and more tests have been added.

Multivariate Kernel weights have been corrected.
gte() now works with a constant covariate.
'itermax' default value has been changed.
Left censored observations can now be inputted by a left endpoints of the censoring interval taking the value NA or 0 (no more 0.0000001). Right censored observations can now be inputted by a right endpoints of the censoring interval taking the value NA or Inf (no more 10000). Events at exact time can now be inputted by equal left and right endpoints of the censoring interval (a very small difference between L and R is no more needed in the input, but the C code still need it to work correctly).
In the data set simul, 0.0000001 values for L and 10000 values for R have have been replaced by NA and for events at exact time L and R are now equals. Also, the data set documentation has been updated and improved.
The package examples have been extended.
The columns of the output value surv have been renamed.
Tests made with testthat have been added in inst/tests (see inst/tests/READMEtest.R).