NEWS.md

In smog: Structural Modeling by using Overlapped Group Penalty

smog 2.1.0

• Use pkgdown to build a package website.
• Added cv.cglasso, plot.cv.cglasso functions for applying a condictional group lasso method to selet biomarkers.
• Added plot.cv.smog function to visualize model selection process by tuning penalty parameters.

smog 2.0

• First version published in CRAN.

smog 1.0

• smog 1.0 is an initial version that contains the work on constrained structural modeling by using overlapped group penalties since June, 2018.

• This version aims to fit a generalized linear model for the cases of small n, large p data, conditioning on the expertise knowledge by non-penalizing some phenotype covariate variables, and penalizing the groups of prognostic and predictive effects by using overlapped L2 and L1 penalties.

smog documentation built on Aug. 10, 2020, 5:07 p.m.