NEWS (original) (raw)

ProSGPV

Penalized Regression with Second-Generation P-Values

ProSGPV 1.0.0

8/6/2021: ProSGPV version 1.0.0 is here!

• Added the reference that proposed the ProSGPV algorithm

ProSGPV 0.3.0

5/4/2021: ProSGPV is faster now!

• Changed the first-stage screening method from cross-validated lasso to one using generalized information criterion
• Removed which.sgpv as there is no variability in selection results given a data set

ProSGPV 0.2.0

3/31/2021: ProSGPV can handle GLM now!

• Added functionalities of variable selection in Logistic, Poisson, and Cox models
• Added the which.sgpv function to find the most frequent model when the two-stage algorithm is used
• Added the S3 plot method for the one-stage algorithm
• Added vignettes to better illustrate how ProSGPV works

ProSGPV 0.1.0

1/6/2021: The first version of ProSGPV is released now!

• Added a NEWS.md file to track changes to the package