bpjandree/autoglm

This is an R package for automated GLM models that can handle large datasets. Supports linear, logit and probit models.

AutoGLM

This is an R package for automated GLM modelling suitable for large datasets. The package resulted from the needs for automated procedures for the statistical calibration of the LUISA model framework that is widely adopted by the European Commision for ex-ante evaluation of policies. Feel free to use it in your own work and contact me with questions.

Current Version

Current version is v.1.0.1b. Reinstall if you're still on v.1.0.0 by executing:

library(devtools)
install_github("BPJandree/AutoGLM")

Release Note for v.1.0.1.b

Edited exportsweightsfile to aligh with geoDMS format. Updating to 1.0.1b will not have any effect on other routines.

Release Note for v.1.0.1.

Release contains several crucial updates:

• Solved an issue with op.t in autoGLM().
• Solved important issues in the opt.h routine. Added a functionality for both single and joint tests.
• Changed wd() as default path to getwd().
• Fixed the warning message: "In if (reclasstable == "default") { :the condition has length > 1 and only the first element will be used".
• The ITdata set has been edited. Several missing values coded with integers have been removed.
• Fixed issues with actions=c(...).
• Added additional functionality for straightforward estimation of logit models using logit(), and integrated the gmm logits within this general command through the option logit(..., method ="gmm").
• gmm estimation through logit() now supports both identity and inverse covariance weighting, and supports a two-step, continously-updated and iteratively-updated estimation methods.
• guessStartVal gmm functionality added, robustified the routines.
• Renamed getCall() to getCall2() to avoid namespace issues with the stats package.
• Updated the example codes.
• Several fixes in the reclassification function.
• Moved most of the miscellaneous functions inside routines that use them to get a cleaner manual.
• simulateLogit now simulates from the correctly specified logit model.
• Several minor bug fixes.

Next Release

Provide me with feedback!

• Update of the manual.
• logit() excepts only a dataframe, will allow for matrix objects to be supplied.

Known bugs (Will be fixed with next update!)

Provide me with feedback!

Example code: You can view some example code here autoGLM_examples.r, or have a look at the vignette.

Feel free to grab the code and contact me if you need help. Just run the following in your R terminal:

library(devtools)
install_github("BPJandree/AutoGLM")