inst/News.md

In rpms: Recursive Partitioning for Modeling Survey Data

----------- Version 0_4_1 ------------------

Allows rpms_forest() an option to run parallel on non-Windows systems

------------ Version 0_4_0 ------------------

Provides function rpms_forest() which fits a random forest model to survey data

Uses loss that is not weighted by proportion of sample - leads to more interpretable splits

Improved version of qtree() function - allows option to show sample size n

-------- Version 0_3_0 --------------

The whole recusive partitioning algorithm is now done in C++

-fixes rounding errors resulting from moving between R and C++

Missing values for categorical variables used for recursive partitioning are treated as their own category

More options added to the qtree() function

Provides a full year of Consumer Expenditure (CE) dataset with more interesting variables

--------- Version 0_2_1 ---------------

Fixes a number of bugs.

Makes splitting on categories faster when e_equ=y~1

Updates variable selection procedure with new perm reps

New version of qtree with more options

---------- Version 0_2_0 ----------

Provide a new function end_nodes()

Fixed bugs in examples using end_nodes()

Included vignette "rpms"

Changed variable selection procedure in case of ties in p-value evaluation

Added labels and caption options to qtree() function

---------- Version 0_1_0 -----------

provides main rpms function with

-two methods: print() and predict()

other functions: in_node(), qtree(), node_plot()

rpms documentation built on June 26, 2021, 1:07 a.m.