-
Notifications
You must be signed in to change notification settings - Fork 40
/
importance.html
263 lines (213 loc) · 11.5 KB
/
importance.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
<!-- Generated by pkgdown: do not edit by hand -->
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Importance
Extract variable importance measures produced by
randomForest and order in decreasing order of
importance. — importance • sl3</title>
<!-- jquery -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/3.4.1/jquery.min.js" integrity="sha256-CSXorXvZcTkaix6Yvo6HppcZGetbYMGWSFlBw8HfCJo=" crossorigin="anonymous"></script>
<!-- Bootstrap -->
<link href="https://cdnjs.cloudflare.com/ajax/libs/bootswatch/3.4.0/flatly/bootstrap.min.css" rel="stylesheet" crossorigin="anonymous" />
<script src="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/3.4.1/js/bootstrap.min.js" integrity="sha256-nuL8/2cJ5NDSSwnKD8VqreErSWHtnEP9E7AySL+1ev4=" crossorigin="anonymous"></script>
<!-- bootstrap-toc -->
<link rel="stylesheet" href="../bootstrap-toc.css">
<script src="../bootstrap-toc.js"></script>
<!-- Font Awesome icons -->
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.12.1/css/all.min.css" integrity="sha256-mmgLkCYLUQbXn0B1SRqzHar6dCnv9oZFPEC1g1cwlkk=" crossorigin="anonymous" />
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.12.1/css/v4-shims.min.css" integrity="sha256-wZjR52fzng1pJHwx4aV2AO3yyTOXrcDW7jBpJtTwVxw=" crossorigin="anonymous" />
<!-- clipboard.js -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/clipboard.js/2.0.6/clipboard.min.js" integrity="sha256-inc5kl9MA1hkeYUt+EC3BhlIgyp/2jDIyBLS6k3UxPI=" crossorigin="anonymous"></script>
<!-- headroom.js -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/headroom/0.11.0/headroom.min.js" integrity="sha256-AsUX4SJE1+yuDu5+mAVzJbuYNPHj/WroHuZ8Ir/CkE0=" crossorigin="anonymous"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/headroom/0.11.0/jQuery.headroom.min.js" integrity="sha256-ZX/yNShbjqsohH1k95liqY9Gd8uOiE1S4vZc+9KQ1K4=" crossorigin="anonymous"></script>
<!-- pkgdown -->
<link href="../pkgdown.css" rel="stylesheet">
<script src="../pkgdown.js"></script>
<meta property="og:title" content="Importance
Extract variable importance measures produced by
randomForest and order in decreasing order of
importance. — importance" />
<meta property="og:description" content="Function that takes a cross-validated fit (i.e., cross-validated learner that
has already been trained on a task), which could be a cross-validated single
learner or super learner, to generate a loss-based variable importance
measure for each predictor, where the predictors are the covariates in the
trained task. This function generates a data.table in which each row
corresponds to the risk difference or risk ratio between the following
two risks: the risk when a predictor is permuted or removed, and the original
risk (i.e., when all predictors are included). A higher risk ratio/difference
corresponds to a more important predictor. A plot can be generated from the
returned data.table by calling companion function
plot_importance." />
<!-- mathjax -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/MathJax.js" integrity="sha256-nvJJv9wWKEm88qvoQl9ekL2J+k/RWIsaSScxxlsrv8k=" crossorigin="anonymous"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/config/TeX-AMS-MML_HTMLorMML.js" integrity="sha256-84DKXVJXs0/F8OTMzX4UR909+jtl4G7SPypPavF+GfA=" crossorigin="anonymous"></script>
<!--[if lt IE 9]>
<script src="https://oss.maxcdn.com/html5shiv/3.7.3/html5shiv.min.js"></script>
<script src="https://oss.maxcdn.com/respond/1.4.2/respond.min.js"></script>
<![endif]-->
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-115145808-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-115145808-1');
</script>
</head>
<body data-spy="scroll" data-target="#toc">
<div class="container template-reference-topic">
<header>
<div class="navbar navbar-default navbar-fixed-top" role="navigation">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#navbar" aria-expanded="false">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<span class="navbar-brand">
<a class="navbar-link" href="../index.html">sl3</a>
<span class="version label label-default" data-toggle="tooltip" data-placement="bottom" title="Released version">1.4.3</span>
</span>
</div>
<div id="navbar" class="navbar-collapse collapse">
<ul class="nav navbar-nav">
<li>
<a href="https://tlverse.org">tlverse</a>
</li>
<li>
<a href="../index.html">sl3</a>
</li>
<li>
<a href="../reference/index.html">Reference</a>
</li>
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown" role="button" aria-expanded="false">
Articles
<span class="caret"></span>
</a>
<ul class="dropdown-menu" role="menu">
<li>
<a href="../articles/intro_sl3.html">Intro to sl3</a>
</li>
<li>
<a href="../articles/SuperLearner_benchmarks.html">sl3 Benchmarks</a>
</li>
<li>
<a href="../slides/sl3-slides.html">sl3 Presentation</a>
</li>
<li>
<a href="../articles/custom_lrnrs.html">Defining Custom sl3 Learners</a>
</li>
</ul>
</li>
</ul>
<ul class="nav navbar-nav navbar-right">
<li>
<a href="https://github.com/tlverse/sl3/">
<span class="fab fa-github fa-lg"></span>
</a>
</li>
</ul>
</div><!--/.nav-collapse -->
</div><!--/.container -->
</div><!--/.navbar -->
</header>
<div class="row">
<div class="col-md-9 contents">
<div class="page-header">
<h1>Importance
Extract variable importance measures produced by
<code><a href='https://rdrr.io/pkg/randomForest/man/randomForest.html'>randomForest</a></code> and order in decreasing order of
importance.</h1>
<small class="dont-index">Source: <a href='https://github.com/tlverse/sl3/blob/master/R/Lrnr_randomForest.R'><code>R/Lrnr_randomForest.R</code></a>, <a href='https://github.com/tlverse/sl3/blob/master/R/importance.R'><code>R/importance.R</code></a></small>
<div class="hidden name"><code>importance.Rd</code></div>
</div>
<div class="ref-description">
<p>Function that takes a cross-validated fit (i.e., cross-validated learner that
has already been trained on a task), which could be a cross-validated single
learner or super learner, to generate a loss-based variable importance
measure for each predictor, where the predictors are the covariates in the
trained task. This function generates a <code>data.table</code> in which each row
corresponds to the risk difference or risk ratio between the following
two risks: the risk when a predictor is permuted or removed, and the original
risk (i.e., when all predictors are included). A higher risk ratio/difference
corresponds to a more important predictor. A plot can be generated from the
returned <code>data.table</code> by calling companion function
<code>plot_importance</code>.</p>
</div>
<pre class="usage"><span class='fu'>importance</span><span class='op'>(</span><span class='va'>fit</span>, loss <span class='op'>=</span> <span class='cn'>NULL</span>, fold_number <span class='op'>=</span> <span class='st'>"validation"</span>,
type <span class='op'>=</span> <span class='fu'><a href='https://rdrr.io/r/base/c.html'>c</a></span><span class='op'>(</span><span class='st'>"remove"</span>, <span class='st'>"permute"</span><span class='op'>)</span>, importance_metric <span class='op'>=</span> <span class='fu'><a href='https://rdrr.io/r/base/c.html'>c</a></span><span class='op'>(</span><span class='st'>"ratio"</span>,
<span class='st'>"difference"</span><span class='op'>)</span><span class='op'>)</span>
<span class='fu'>importance</span><span class='op'>(</span><span class='va'>fit</span>, loss <span class='op'>=</span> <span class='cn'>NULL</span>, fold_number <span class='op'>=</span> <span class='st'>"validation"</span>,
type <span class='op'>=</span> <span class='fu'><a href='https://rdrr.io/r/base/c.html'>c</a></span><span class='op'>(</span><span class='st'>"remove"</span>, <span class='st'>"permute"</span><span class='op'>)</span>, importance_metric <span class='op'>=</span> <span class='fu'><a href='https://rdrr.io/r/base/c.html'>c</a></span><span class='op'>(</span><span class='st'>"ratio"</span>,
<span class='st'>"difference"</span><span class='op'>)</span><span class='op'>)</span></pre>
<h2 class="hasAnchor" id="arguments"><a class="anchor" href="#arguments"></a>Arguments</h2>
<table class="ref-arguments">
<colgroup><col class="name" /><col class="desc" /></colgroup>
<tr>
<th>fit</th>
<td><p>A trained cross-validated learner (e.g., cv stack, super learner),
from which cross-validated predictions can be generated.</p></td>
</tr>
<tr>
<th>loss</th>
<td><p>The loss function for evaluating the risk. Defaults according to
outcome type: squared error loss for continuous outcomes, and negative
log-likelihood loss for discrete outcomes. See <code><a href='loss_functions.html'>loss_functions</a></code>.</p></td>
</tr>
<tr>
<th>fold_number</th>
<td><p>The fold number to use for obtaining the predictions
from the fit. Either a positive integer for obtaining predictions from a
specific fold's fit; <code>"full"</code> for obtaining predictions from a fit on
all of the data, or <code>"validation"</code> (default) for obtaining
cross-validated predictions, where the data used for training and prediction
never overlaps across the folds. Note that if a positive integer or
<code>"full"</code> is supplied, then there will be overlap between the data
used for training and prediction.</p></td>
</tr>
<tr>
<th>type</th>
<td><p>Which method should be used to obscure the relationship between
the covariate and the outcome. When type is <code>"remove"</code> (default), each
covariate is removed from the task and the cross-validated learner is refit
to this modified task and then predictions are obtained from this refit.
When type is <code>"permute"</code>, each covariate is permuted (sampled without
replacement), and then predictions are obtained from this modified data with
one permuted covariate.</p></td>
</tr>
<tr>
<th>importance_metric</th>
<td><p>Either <code>"ratio"</code> (default), which returns
the risk with the permuted/removed X divided by observed risk with all X; or
<code>"difference"</code>, which returns the difference between the risk with the
permuted/removed X and the observed risk.</p></td>
</tr>
</table>
<h2 class="hasAnchor" id="value"><a class="anchor" href="#value"></a>Value</h2>
<p>A <code>data.table</code> of variable importance for each covariate.</p>
</div>
<div class="col-md-3 hidden-xs hidden-sm" id="pkgdown-sidebar">
<nav id="toc" data-toggle="toc" class="sticky-top">
<h2 data-toc-skip>Contents</h2>
</nav>
</div>
</div>
<footer>
<div class="copyright">
<p>Developed by Jeremy Coyle, Nima Hejazi, Oleg Sofrygin, Ivana Malenica, Rachael Phillips.</p>
</div>
<div class="pkgdown">
<p>Site built with <a href="https://pkgdown.r-lib.org/">pkgdown</a> 1.6.1.</p>
</div>
</footer>
</div>
</body>
</html>