- Fix log-likelihood in
- Enable further input arguments for legend in plots of residuals.
- Implementation of residuals: Cox-Snell, randomized quantile residuals and martingale residuals.
- We fixed the hazard function routine implemented to compute hazard functions for any distribution.
- We also implemented a routine to compute cumulative hazard functions for any distribution.
- We fixed and updated the documentation.
CRAN release: 2022-12-10
- We changed the behaviour of
plot.HazardShapefunction, its responsibility was splitted. Now, this function just plots the empirical TTT function, and we also created the
legend.HazardShapeto customize the legend options.
- We fixed some examples in the documentation.
- EstimationTools was build in R 4.2.1.
- From this version,
maxlogLregreturns the censorship matrix. Check the documentation for further information about the censorship matrix.
- We have implemented a wrapper (an API) of integration routines.
- New!!! genetic algorithm (
GA::ga) implemented in
- Custom optimization interface enabled (tests required).
- Link functions updated, precision problems solved.
- Link functions specified in summary (maxlogLreg) and print methods.
- Bug in answer class object in
TTT_hazard_shapewas implemented as method for classes
Hazard_Shapefunction deprecated. We implemented a ‘print’ method for
plot.HazardShape: More flexibility in plot options enabled.
- Bug in
se.fit(predict method) fixed.
- Default value for data (NULL) in order to handle data inside another function (use with
applyfamily) in TTT functions.
- New data sets and examples for them added. They were also referenced.
CRAN release: 2021-03-10
- Function for hazard shape estimation implemented. A plot method was also implemented.
- Bug fixed. Formula handle with survival objects without factor classification in TTTE_Analytic is fixed.
CRAN release: 2020-08-03
EstimationTools was build in R 4.0.2.
We have implemented non-centrality parameter detection. This parameter must be named
ncpin any custom distribution.
We have implemented computation and plots of empirical total time on test (TTT).
Standard error computation is now performed inside
Standard errror value is now stored in the list case named ‘fit’. It can be called typing
objectname$fit$StdE, where ‘objectname’ is the object where the model is stored.
The new method (function)
bootstrap_maxlogLwas created for standard error computation through bootstrap.
‘summary’ method only reports results. It does not perform any computation, bootstrap through ‘summary’ function has been deprecated.
We have made a careful spelling check in our documentation.