pp_ggparty

projection pursuit regression tree plot with independent variable

tree

It was developed as a tool for exploring 'PPTreereg' (Projection Pursuit TREE of REGression).
It uses various projection pursuit indexes and 'XAI' (eXplainable Artificial Intelligence) methods to help
understand the model by finding connections between the input variables and prediction values of the model.
The 'KernelSHAP' (Aas, Jullum and Løland (2019) <arXiv:1903.10464>) algorithm was modified to fit ‘PPTreereg’,
and some codes were modified from the 'shapr' package (Sellereite, Nikolai, and Martin Jullum (2020) <doi:10.21105/joss.02027>).
The implemented methods help to explore the model at the single instance level as well as at the whole dataset level.
Users can compare with other machine learning models by applying it to the 'DALEX' package of 'R'.

HyunSun Cho

PPtreeregViz

Projection Pursuit Regression Tree Visualization

Eun-Kyung Lee

Nikolai Sellereite

Martin Jullum

Annabelle Redelmeier

Norsk Regnesentral 

pp_ggparty function

<dl><dt>PPTreeregOBJ</dt>
<dd>PPTreereg class object</dd>
<dt>ind_variable</dt>
<dd>independent variable to show</dd>
<dt>final.rule</dt>
<dd>final rule to assign numerical values in the final nodes.
1: mean value in the final nodes
2: median value in the final nodes
3: using optimal projection
4: using all independent variables
5: using several significant independent variables</dd>
<dt>Rule</dt>
<dd>split rule
1: mean of two group means
2: weighted mean of two group means - weight with group size
3: weighted mean of two group means - weight with group sd
4: weighted mean of two group means - weight with group se
5: mean of two group medians
6: weighted mean of two group medians - weight with group size
7: weighted mean of two group median - weight with group IQR
8: weighted mean of two group median - weight with group IQR
 and group size</dd>
<dt>...</dt>
<dd>arguments to be passed to methods</dd></dl>

Arguments

PPTreereg plot with independent variable — pp_ggparty

<dl>

<dt>PPTreeregOBJ</dt>
<dd>PPTreereg class object</dd>


<dt>ind_variable</dt>
<dd>independent variable to show</dd>


<dt>final.rule</dt>
<dd>final rule to assign numerical values in the final nodes.
1: mean value in the final nodes
2: median value in the final nodes
3: using optimal projection
4: using all independent variables
5: using several significant independent variables</dd>


<dt>Rule</dt>
<dd>split rule
1: mean of two group means
2: weighted mean of two group means - weight with group size
3: weighted mean of two group means - weight with group sd
4: weighted mean of two group means - weight with group se
5: mean of two group medians
6: weighted mean of two group medians - weight with group size
7: weighted mean of two group median - weight with group IQR
8: weighted mean of two group median - weight with group IQR
 and group size</dd>


<dt>...</dt>
<dd>arguments to be passed to methods</dd>

</dl>

pp_ggparty: PPTreereg plot with independent variable

Description

Usage

Value

Arguments

Details

Examples