Quick start for `abess`: Linear regression
Classification: Logistic Regression and Multinomial Extension
Positive response: Poisson and Gamma regression
Best Subset Selection for Censored Response
Multi-Response Linear Regression
Principal component analysis
Tips for faster computation
ABESS algorithm: details
Power of abess
Robust Principal Component Analysis
Allow the other criterion for model selection: AUC for (multinomial) logistic regression such as the area under the curve (AUC).
Simplify the C++ code structure.
Fix note “Specified C++11: please update to current default of C++17” in CRAN.
Adapt to the API change of the
Change the package structure such that the API functions can reuse the utility function. It facilitates the testing for package.
Update citation information.
Support generalized linear model for ordinal response, also named as rank learning in machine learning community.
Support robust principal analysis
Modify R package structure to make many internal components are reusable.
Support generalized linear model when the link function is Gamma distribution. By setting
family = "gamma"
function, users can analyze the dataset with a positive valued and skewed response.
Support flexible support size for sequential principal component analysis (PCA), accompanied with several helpful generic function like
Support user-specified cross validation division for
function by additional argument
Support robust principal component analysis now. A new R function
can access it.
Improve the R package document by: adding more details and giving more links related to core functions.
Add docs2search for R’s website
Support important searching to improve computational efficiency when dimension is 10,000.
Support sparse matrix as input
Support golden section search for optimal support size
Support ridge-regularized penalty as a generic component
Support group subset selection as a generic component
Best subset selection for principal component analysis via
Initial stable version abess package
Support best subset selection for linear regression, logistic regression, poisson regression, cox proportional hazard regression, multi-gaussian regression, multi-nominal regression.
Support nuisance selection as a generic component
Unified API for cross validation and information criterion to select the optimal support size.
A documentation website is support for