CHANGES in `gzippath' VERSION 0.1_7 (2012-8-20)
o change glmcd on how the response variable y is scaled for family other than gaussian
o 8-25-2012 modification on gzipath_0.1-9.tar.gz
b0 should be the average of response variable y for non-standardized case, in an agreement with the so determined lambda max for which beta = 0. This is correct in R function lmnet, see zipath_0.1-2_08_10_2012.tar.gz
CHANGES in `gzipath' VERSION 0.1_13 (2013-11-24)
o for weighted penalized regression glmreg_fit, with standardize=TRUE, the weights are normalized to n*weights/sum(weights). Now added standardize=FALSE and the weights are not normalized.
o zipath function was modified to incorporate glmreg_fit for weighted penalized Poisson and Negative Binomial regression. The tuning parameters were changed based on modified glmreg_fit.
CHANGES in `gzipath' VERSION 0.1_14 (2014-5-8)
o for rescale=TRUE and penalty="mnet" or "snet" in zipath, data driven lambda.zero[1] (i.e., the maximum value of lambda in zero component) is too small, now changed to 0.5*sqrt(lambda.zero[1]). This may be modified if the maximum value of lambda in glmreg was changed in the similar situations.
CHANGES in `gzipath' VERSION 0.1_14 (2014-6-9)
o change package name to `mpath'
CHANGES in `mpath' VERSION 0.1_15 (2014-9-23)
o apply patch from Martyn Plummer to replace isna and lgamma in src Fortran file to avoid failure on some platforms
CHANGES in `mpath' VERSION 0.1_16 (2014-10-25)
o change src/midloop.f, now saturated model only applies to logistic regression
o add penalty.factor.count and penalty.factor.zero in zipath function to allow different penalties for each of the variables.
o add vignette to illustrate analysis of German health care demand
CHANGES in `mpath' VERSION 0.1_17 (2014-11-10)
o fix a bug in function estfunReg, which is supposed to work only for one penalty parameter
o fix a bug in src/lmnetGaus.f, computing meanx
CHANGES in `mpath' VERSION 0.1_18 (2015-2-26)
o vignette file was updated to be consistent with the corresponding paper Wang et al (2015)
o fix bugs in cv.glmreg.Rd, cv.glmregNB.Rd and cv.zipath.Rd
o in findlam function, verify 0 < alpha <= 1
o fix a bug with intercept estimate for family="gaussian"
o remove the threshold value mu=1000 for family=4 in midloop.f, lines 167 and 207
o change function name predict.zeroinfl1 to predictzeroinfl1
CHANGES in `mpath' VERSION 0.1_19 (2015-7-23)
o editorial changes to files in man directory
CHANGES in `mpath' VERSION 0.1_20 (2015-7-23)
o rebuild static_german.pdf, updating a reference
o fix a bug in loss function for family="gaussian"
CHANGES in `mpath' VERSION 0.1_21 (2015-11-23)
o minor changes to print lambda 4 digits
CHANGES in `mpath' VERSION 0.2-1 (12-09-2015)
o add parallel computing for cross-validation in cross-validation function
o remove vuong test, based on Wilson (2015), The misuse of the Vuong test for non-nested models to test for
zero-inflation, Economics Letters, 127: 51:53
CHANGES in `mpath' VERSION 0.2-3 (08-10-2016)
o change the format of b0 and beta in glmreg function when standardize=FALSE, to be consistent with standardize=TRUE
o for findlam function in glmreg.R, change
if(standardize)
weights <- weights/sum(weights)
to
if(family=="gaussian" || standardize)
weights <- weights/sum(weights)
This corrects an error in glmreg with family="gaussian" and standardize=FALSE
o In src/lmnetGLM.f, remove lmnetGLM subroutine, which was not used anymore
o fix subroutines in lmnetGLM.f, lmnetGaussian.f, midloop.f and outloop.f regarding gfortran warning messages with an email request from cran.R-project on August 28, 2016
CHANGES in `mpath' VERSION 0.2-4 (08-30-2016)
o after a change of penGLM subroutine in midloop.f (removing unused argument n), glmreg.R and zipath.R were changed when invoking penGLM.
CHANGES in `mpath' VERSION 0.3-1 (09-16-2016)
o add ncl and nclreg, and relevant files for nonconvex-based models
o fix a bug computing penval in glmreg, add penalty.factor when calling penGLM in glmreg and zipath. These bugs don't affect the previous analysis results
CHANGES in `mpath' VERSION 0.3-2 (07-14-2017)
o To register native routines and to disable symbol search.
$cd mpath
$R
> tools::package_native_routine_registration_skeleton(".")
copy and paste the full output in a packagename_init.c (i.e., mpath_init.c) file to be put in src/
o change convergence criteria in nclreg.R
d1 <- sum((fk_old - fk)^2)
#d1 <- sum((fk_old - fk)^2)/sum(fk_old^2) ### this can cause a problem if fk_old is zero
o another change of convergence criteria in nclreg.R: comment out the following lines
#if(k > 1){
# if(los[k] > los[k-1]){
#k <- iter
#}
#}
o fix a bug in nclreg.R: lambda.min.ratio was not used in computing lambda with glmreg_fit
o in nclreg.R, computing solution path stops if the numbers in exponential function become too big (infinite)
if(any(is.nan(h))){ # exit loop
stopit <- TRUE
break
}
o add type.init and relevant arguments in nclreg_fit function to have three types of initial parameters and lambda values
o replace fraction with lambda in cv.glmreg.R, cv.nclreg.R, cv.zipath.R and cv.glmregNB.R. This only affects the values of x-axis in plots, but not cross-validation results
o add type.path to have two different updating methods when computing solution paths
CHANGES in `mpath' VERSION 0.3-3 (08-17-2017)
o cycle through active set with modified codes in Fortran files in src directory
CHANGES in `mpath' VERSION 0.3-4 (08-22-2017)
o add type="active" in R/nclreg_fit function
o fix a bug with cv.zipath when plot.it=TRUE
o add stopImplicitCluster() after registerDoParallel
CHANGES in `mpath' VERSION 0.3-5 (10-17-2017)
o add vignette to illustrate classification of cancer patients with penalized robust nonconvex loss functions
o change zipath function to cycle through warm start. The results can be different from the previous versions.
o fix bugs with zipath when trace=TRUE and the zero component is an intercept-only model
CHANGES in `mpath' VERSION 0.3-6 (10-04-2018)
o fix bugs when all elements of penalty.factor=0 in glmreg_fit or glmregNB. In this case, we set lambda=0 for each variable, and all penalty.factor=1.
o fix bugs when all elements of penalty.factor.count=0 or penalty.factor.zero=0 in zipath. In this case, we set corresponding lambda = 0.
o fix bugs when activeset=NULL in nclreg function
CHANGES in `mpath' VERSION 0.3-7 (01-26-2019)
o changes were made to make offset in glmreg work. This includes changes in outloop.f, midloop.f, midloopGLM.f, lmnetGLM.
o Changes were made in midloopGLM.f so that iterations run and check for convergency for b0 as well as beta
o Changes were made in zipath to make offset work
o offset is now working in glmregNB, zipath, cv.glmreg, cv.glmregNB and cv.zipath
CHANGES in `mpath' VERSION 0.3-8 (03-03-2019)
o R zipath function has both Fortran and R computations, can be useful to compare the results.
o changes were made on how active set works in src/outloop.f file. Now, we cycle through the active set until convergence, then cycle through all the variables ONCE, this is an reverse to the previous procedure. This includes changes in outloop.f, midloop.f, midloopGLM.f, lmnetGLM, add src/find_activeset.f to simplify the codes.
o R/nclreg seems to be incorrect when update beta[activeset, i], since activeset is relative to the current x.act, not x. The new Fortran subroutine using dynamic allocation to deal with this. The Fortran subroutine nclreg_fortran.f conducts the three layers of iteration loops, lambda, MM and internal penalty estimation loop.
o Not a good idea to convert theta.ml R function to Fortran subroutine since the speed actually is slower in Fortran. This is probably because digamma and trigamma functions in a Fortran subroutine used C functions. In contrast, the R function theta.ml directly used these C functions. Maybe there are other reasons. But we leave a glmregNB.R with the Fortran subroutine glmregnb_fortran.
CHANGES in `mpath' VERSION 0.3-9 (04-06-2019)
o Major improvement on zipath.R. Replace the R codes with Fortran codes to compute lmax. Compute the solution path with zipath_fortran subroutine to replace the R codes.
CHANGES in `mpath' VERSION 0.3-10 (04-12-2019)
o update zipath_fortran subroutine with an active set approach. Is this faster than nonactive approach?
CHANGES in `mpath' VERSION 0.3-11 (04-12-2019)
o implement both active and nonactive path in R zipath function
CHANGES in `mpath' VERSION 0.3-13 (05-18-2019)
o fix bugs in Fortran subroutines, mostly related to uninitialized values after dynamic allocate of vectors.
o in vignettes/brcancer.Rnw, compare type.path="nonative" and "active"
o in man/nclreg.Rd, provide examples of different solution paths
CHANGES in `mpath' VERSION 0.3-14 (05-24-2019)
o zipath can be slow for some data, and change outloop subroutine from active set to nonactive set can help substantially. The reason is that the active set approach can repeat the cycle without convergence. Will be addressed in version 0.3-15.
CHANGES in `mpath' VERSION 0.3-15 (06-03-2019)
o still use active set, as in version 0.3-13, but change the number of maximum iteration from 100 to 2 when determining active set in outloop.f. Similarly in lmnetGaussian.f. This will reduce the computing times in case of nonconvergence. The program looks for a cycle that repeats itself, default is 2
o fix bugs in nclreg_ad.f regarding m_act and jk, and similar bugs in zipath_active.f
CHANGES in `mpath' VERSION 0.3-16 (07-09-2019)
o to avoid mismatch between predictor variables and the estimated coefficients - this can happen if the input data contains response variable but not appropriately handled, add the following line in R/predict.glmreg.R
if(dim(object$beta)[1] != m) stop("number of columns in newx should be the same as number of rows in beta\n")