Movatterモバイル変換

cjbart 0.3.2

• Update documentation: correct details line forcjbart().
• Add compatibility so passing seed argument on Windows returnswarning rather than error
• Add unit tests for OS compatibility

cjbart 0.3.1

• Fix documentation: to plot the variable importance grid, thex argument takes the result of callinghet_vimp() and notIMCE() as previouslydocumented (in v0.3.0)

• Update package dependencies

cjbart 0.3.0

• Rename arguments inhet_vimp() frommodeltoimces andoutcomes tolevelsfor consistency across functions
• Changehet_vimp() so it returns a list, including theresults and an attribute-level look-up table
• Added ability to runhet_vimp() calculation inparallel
• Fixed bug when predicting continuous conjoint outcomes
• Add newpIMCE() function to calculate populationindividual-level marginal component effects
• Add newAMCE() function that includes Bayesianuncertainty estimate
• Add check for unique levels across attributes. If not, theattribute-level values are modified to ensure uniqueness
• Add functionality to improve quality of outputs
• Update the vignette to reflect new functionality
• Add ability to drop attributes from IMCE analysis
• Remove dependencies not needed

cjbart 0.2.2

• Re-release after dependencies back on CRAN

cjbart 0.2.1

• Minor internal changes to keep pace with R development
• Minor changes to vignette wording

cjbart 0.2.0

• Addresses unbalanced chunk delimiters in package vignette
• Adds new functionality for analyzing heterogeneity after IMCEprediction using random forest variable importance measures

cjbart 0.1.0

• First release including all core functionality
• Includes wrapper of BART estimation, IMCE prediction, andgenerics