2.0-1
- correct the value for 'Selected.alpha' in the documentation of the NPCoImp
- modify an error message
- update references
2.0-0
- added a new function (class NPCoImp) that implements an additional imputation method based on the empirical conditional copula.
1.0
- enlarge the set of copula models: t-Student with df.fixed=FALSE, rotated copulas, unstructured normal copulas
- fixed a problem with the starting values of the copulas to be estimated
- fixed some problems with the extraction of the class according to the kind of copula model
0.3-1
- fixed a bug on x.up and x.lo arguments of the CoImp function;
- improved the q.up (x.up) and q.lo (x.lo) arguments of the CoImp function: now it is possible to specify a different value for each margin;
- improved the MAR function.
0.3-0
- introduced some arguments in the CoImp function used in the Hit or Miss and in the maximization of each marginal density function;
- fixed a problem in the MCAR and MAR functions (missed a set.seed));
- improved the argument plot of the CoImp function and the colnames in the two plots;
- added some checks in all the functions;
- updated references in the documentation;
- changed the loglikCopula function call according to the changes in the copula package;
- fixed a problem in the second while of the hitormiss function ('and' was substituted by 'or');
- added a check on the presence of at least one complete record in the CoImp function;
- added a check on the matrix in input of the functions MAR and MCAR.
0.2-3
- cleaned and improved the documentation;
- moved almost all the dependencies to the Imports section;
- removed the dependence on tseriesEntropy package which is not on the CRAN.
0.2-2
- removed the dependence on cubature package;
- changed the function fcond.mod in CoImp_Esterne.R (the adaptIntegrate function of the cubature package is not used anymore
and the check due to a bug in dcopula function have been deleted);
- hitormiss function does not use anymore the vectorized version of the fcond function;
- modified the computation of the median values of the data;
- reintroduced methods plot and show;
- introduced a function (MCAR) that generates missing completely at random;
- introduced a function (MAR) that generates missing at random through a logistic multinomial regression model;
- introduced a function (PerfMeasure) that computes a set of performance measures for a given imputed data set;
- deleted the function for multiple imputation (CoImpEval).
0.2-1
- added the cubature package in the Description file;
- changed the function fcond.mod in CoImp_Esterne.R: now it uses the adaptIntegrate function of the cubature package;
- example in CoImp.Rd: deleted three TRUE values in the call of the CoImp function (TRUE is default).
0.1-2
- packaged under R 2.15.0;
- deleted methods plot and show: problems should be fixed.
0.1-1
- added a check to fcond.mod related to a bug in dcopula (package copula); to be removed;
- fixed import warning on import(graphics).