NEWS.md

In Jurian/bump.hunting: Subgroup Discovery and Bump Hunting

subgroup.discovery v0.3.0 (Release date: 2020-01-30)

Changes:

• Moved all computation to c++ code, resulting package is now much faster
• Finding the optimal peel is done in a parallel map-reduce fashion
• There is a dedicated function called prim.data.prepare() to facilitate the proper setup of input data
• The covering strategies have been removed, package is now in line with common practices, i.e. you build a model using prim() and then evaluate it using predict().
• To find observations which fall into a certain box, use the prim.box.index() function

subgroup.discovery v0.3.1 (Release date: 2020-02-06)

Changes:

• Removed all namespaces except Rcpp to comply with best practices (especially not using std)
• Changed the pattern of sub-box creation so at no point there is a uninstantiated value

subgroup.discovery v0.3.2 (Release date: 2020-03-18)

Changes:

• The minimal R version has been set to 3.0.0, so this package can also be tested on some architectures on CRAN which use an older version
• The issue with ASAN / UBSAN warnings in CRAN submissions should be resolved thanks to update 5.0.0 from RcppParallel package

Jurian/bump.hunting documentation built on March 24, 2020, 8:07 a.m.