gauss_cat_sampler_random

The <code>gauss_cat_sampler_random</code> generates random samples from the generative distribution defined by the
output of the vaeac. The random sample is generated by sampling from the inferred Gaussian and categorical
distributions for the continuous and categorical features, respectively.

internal

Complex machine learning models are often hard to interpret. However, in
many situations it is crucial to understand and explain why a model made a specific
prediction. Shapley values is the only method for such prediction explanation framework
with a solid theoretical foundation. Previously known methods for estimating the Shapley
values do, however, assume feature independence. This package implements methods which accounts for any feature
dependence, and thereby produces more accurate estimates of the true Shapley values.
An accompanying 'Python' wrapper ('shaprpy') is available through the GitHub repository.

Martin Jullum

shapr

Prediction Explanation with Dependence-Aware Shapley Values

Lars Henry Berge Olsen

Annabelle Redelmeier

Jon Lachmann

Nikolai Sellereite

Anders Løland

Jens Christian Wahl

Camilla Lingjærde

Norsk Regnesentral 

gauss_cat_sampler_random function

<dl><dt>one_hot_max_sizes</dt>
<dd>A torch tensor of dimension <code>n_features</code> containing the one hot sizes of the <code>n_features</code>
features. That is, if the <code>i</code>th feature is a categorical feature with 5 levels, then <code>one_hot_max_sizes[i] = 5</code>.
While the size for continuous features can either be <code>0</code> or <code>1</code>.</dd>
<dt>min_sigma</dt>
<dd>For stability it might be desirable that the minimal sigma is not too close to zero.</dd>
<dt>min_prob</dt>
<dd>For stability it might be desirable that the minimal probability is not too close to zero.</dd></dl>

Arguments

Author

A torch::nn_module() Representing a gauss_cat_sampler_random — gauss_cat_sampler_random

<dl>

<dt>one_hot_max_sizes</dt>
<dd>A torch tensor of dimension <code>n_features</code> containing the one hot sizes of the <code>n_features</code>
features. That is, if the <code>i</code>th feature is a categorical feature with 5 levels, then <code>one_hot_max_sizes[i] = 5</code>.
While the size for continuous features can either be <code>0</code> or <code>1</code>.</dd>


<dt>min_sigma</dt>
<dd>For stability it might be desirable that the minimal sigma is not too close to zero.</dd>


<dt>min_prob</dt>
<dd>For stability it might be desirable that the minimal probability is not too close to zero.</dd>

</dl>

A <code>torch::nn_module()</code> Representing a gauss_cat_sampler_random

gauss_cat_sampler_random: A `torch::nn_module()` Representing a gauss_cat_sampler_random

Description

Usage

Arguments

Author