P281424/AMR
1
0
Fork 0
mirror of https://github.com/msberends/AMR.git synced 2025-07-27 12:27:56 +02:00
Code Releases Activity

Built site for AMR@3.0.0.9004: 1013ef6

Browse Source
This commit is contained in:
github-actions
2025-06-13 15:13:34 +00:00
parent 327130f5b6
commit bf7668e26f
110 changed files with 965 additions and 267 deletions
Download Patch File Download Diff File Expand all files Collapse all files

298
articles/AMR_with_tidymodels.html
View File

@ -30,7 +30,7 @@
<a class="navbar-brand me-2" href="../index.html">AMR (for R)</a>
<small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">3.0.0.9002</small>
<small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">3.0.0.9004</small>
<button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@ -101,6 +101,12 @@ functions like <code><a href="../reference/antimicrobial_selectors.html">aminogl
<p>In this post, we will explore how to use the <code>tidymodels</code>
framework to predict resistance patterns in the
<code>example_isolates</code> dataset in two examples.</p>
<p>This post contains the following examples:</p>
<ol style="list-style-type: decimal">
<li>Using Antimicrobial Selectors</li>
<li>Predicting ESBL Presence Using Raw MICs</li>
<li>Predicting AMR Over Time</li>
</ol>
<div class="section level2">
<h2 id="example-1-using-antimicrobial-selectors">Example 1: Using Antimicrobial Selectors<a class="anchor" aria-label="anchor" href="#example-1-using-antimicrobial-selectors"></a>
</h2>
@ -426,17 +432,255 @@ and reproducibly.</p>
</div>
</div>
<div class="section level2">
<h2 id="example-2-predicting-amr-over-time">Example 2: Predicting AMR Over Time<a class="anchor" aria-label="anchor" href="#example-2-predicting-amr-over-time"></a>
<h2 id="example-2-predicting-esbl-presence-using-raw-mics">Example 2: Predicting ESBL Presence Using Raw MICs<a class="anchor" aria-label="anchor" href="#example-2-predicting-esbl-presence-using-raw-mics"></a>
</h2>
<p>In this second example, we aim to predict antimicrobial resistance
<p>In this second example, we demonstrate how to use
<code>&lt;mic&gt;</code> columns directly in <code>tidymodels</code>
workflows using AMR-specific recipe steps. This includes a
transformation to <code>log2</code> scale using
<code><a href="../reference/amr-tidymodels.html">step_mic_log2()</a></code>, which prepares MIC values for use in
classification models.</p>
<p>This approach and idea formed the basis for the publication <a href="https://doi.org/10.3389/fmicb.2025.1582703" class="external-link">DOI:
10.3389/fmicb.2025.1582703</a> to model the presence of
extended-spectrum beta-lactamases (ESBL).</p>
<div class="section level3">
<h3 id="objective-1">
<strong>Objective</strong><a class="anchor" aria-label="anchor" href="#objective-1"></a>
</h3>
<p>Our goal is to:</p>
<ol style="list-style-type: decimal">
<li>Use raw MIC values to predict whether a bacterial isolate produces
ESBL.</li>
<li>Apply AMR-aware preprocessing in a <code>tidymodels</code>
recipe.</li>
<li>Train a classification model and evaluate its predictive
performance.</li>
</ol>
</div>
<div class="section level3">
<h3 id="data-preparation-1">
<strong>Data Preparation</strong><a class="anchor" aria-label="anchor" href="#data-preparation-1"></a>
</h3>
<p>We use the <code>esbl_isolates</code> dataset that comes with the AMR
package.</p>
<div class="sourceCode" id="cb10"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Load required libraries</span></span>
<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://amr-for-r.org">AMR</a></span><span class="op">)</span></span>
<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://tidymodels.tidymodels.org" class="external-link">tidymodels</a></span><span class="op">)</span></span>
<span></span>
<span><span class="co"># View the esbl_isolates data set</span></span>
<span><span class="va">esbl_isolates</span></span>
<span><span class="co">#&gt; <span style="color: #949494;"># A tibble: 500 × 19</span></span></span>
<span><span class="co">#&gt; esbl genus AMC AMP TZP CXM FOX CTX CAZ GEN TOB TMP SXT</span></span>
<span><span class="co">#&gt; <span style="color: #949494; font-style: italic;">&lt;lgl&gt;</span> <span style="color: #949494; font-style: italic;">&lt;chr&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span> <span style="color: #949494; font-style: italic;">&lt;mic&gt;</span></span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 1</span> FALSE Esch… 32 32 4 64 64 8<span style="color: #BBBBBB;">.00</span> 8<span style="color: #BBBBBB;">.00</span> 1 1 16<span style="color: #BBBBBB;">.0</span> 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 2</span> FALSE Esch… 32 32 4 64 64 4<span style="color: #BBBBBB;">.00</span> 8<span style="color: #BBBBBB;">.00</span> 1 1 16<span style="color: #BBBBBB;">.0</span> 320</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 3</span> FALSE Esch… 4 2 64 8 4 8<span style="color: #BBBBBB;">.00</span> 0.12 16 16 0.5 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 4</span> FALSE Kleb… 32 32 16 64 64 8<span style="color: #BBBBBB;">.00</span> 8<span style="color: #BBBBBB;">.00</span> 1 1 0.5 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 5</span> FALSE Esch… 32 32 4 4 4 0.25 2<span style="color: #BBBBBB;">.00</span> 1 1 16<span style="color: #BBBBBB;">.0</span> 320</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 6</span> FALSE Citr… 32 32 16 64 64 64<span style="color: #BBBBBB;">.00</span> 32<span style="color: #BBBBBB;">.00</span> 1 1 0.5 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 7</span> FALSE Morg… 32 32 4 64 64 16<span style="color: #BBBBBB;">.00</span> 2<span style="color: #BBBBBB;">.00</span> 1 1 0.5 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 8</span> FALSE Prot… 16 32 4 1 4 8<span style="color: #BBBBBB;">.00</span> 0.12 1 1 16<span style="color: #BBBBBB;">.0</span> 320</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;"> 9</span> FALSE Ente… 32 32 8 64 64 32<span style="color: #BBBBBB;">.00</span> 4<span style="color: #BBBBBB;">.00</span> 1 1 0.5 20</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;">10</span> FALSE Citr… 32 32 32 64 64 8<span style="color: #BBBBBB;">.00</span> 64<span style="color: #BBBBBB;">.00</span> 1 1 16<span style="color: #BBBBBB;">.0</span> 320</span></span>
<span><span class="co">#&gt; <span style="color: #949494;"># ℹ 490 more rows</span></span></span>
<span><span class="co">#&gt; <span style="color: #949494;"># ℹ 6 more variables: NIT &lt;mic&gt;, FOS &lt;mic&gt;, CIP &lt;mic&gt;, IPM &lt;mic&gt;, MEM &lt;mic&gt;,</span></span></span>
<span><span class="co">#&gt; <span style="color: #949494;"># COL &lt;mic&gt;</span></span></span>
<span></span>
<span><span class="co"># Prepare a binary outcome and convert to ordered factor</span></span>
<span><span class="va">data</span> <span class="op">&lt;-</span> <span class="va">esbl_isolates</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span>esbl <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/factor.html" class="external-link">factor</a></span><span class="op">(</span><span class="va">esbl</span>, levels <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span><span class="cn">FALSE</span>, <span class="cn">TRUE</span><span class="op">)</span>, ordered <span class="op">=</span> <span class="cn">TRUE</span><span class="op">)</span><span class="op">)</span></span></code></pre></div>
<p><strong>Explanation:</strong></p>
<ul>
<li>
<code>esbl_isolates</code>: Contains MIC test results and ESBL
status for each isolate.</li>
<li>
<code>mutate(esbl = ...)</code>: Converts the target column to an
ordered factor for classification.</li>
</ul>
</div>
<div class="section level3">
<h3 id="defining-the-workflow-1">
<strong>Defining the Workflow</strong><a class="anchor" aria-label="anchor" href="#defining-the-workflow-1"></a>
</h3>
<div class="section level4">
<h4 id="preprocessing-with-a-recipe-1">1. Preprocessing with a Recipe<a class="anchor" aria-label="anchor" href="#preprocessing-with-a-recipe-1"></a>
</h4>
<p>We use our <code><a href="../reference/amr-tidymodels.html">step_mic_log2()</a></code> function to log2-transform
MIC values, ensuring that MICs are numeric and properly scaled. All MIC
predictors can easily and agnostically selected using the new
<code><a href="../reference/amr-tidymodels.html">all_mic_predictors()</a></code>:</p>
<div class="sourceCode" id="cb11"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Split into training and testing sets</span></span>
<span><span class="fu"><a href="https://rdrr.io/r/base/Random.html" class="external-link">set.seed</a></span><span class="op">(</span><span class="fl">123</span><span class="op">)</span></span>
<span><span class="va">split</span> <span class="op">&lt;-</span> <span class="fu">initial_split</span><span class="op">(</span><span class="va">data</span><span class="op">)</span></span>
<span><span class="va">training_data</span> <span class="op">&lt;-</span> <span class="fu">training</span><span class="op">(</span><span class="va">split</span><span class="op">)</span></span>
<span><span class="va">testing_data</span> <span class="op">&lt;-</span> <span class="fu">testing</span><span class="op">(</span><span class="va">split</span><span class="op">)</span></span>
<span></span>
<span><span class="co"># Define the recipe</span></span>
<span><span class="va">mic_recipe</span> <span class="op">&lt;-</span> <span class="fu">recipe</span><span class="op">(</span><span class="va">esbl</span> <span class="op">~</span> <span class="va">.</span>, data <span class="op">=</span> <span class="va">training_data</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">remove_role</span><span class="op">(</span><span class="va">genus</span>, old_role <span class="op">=</span> <span class="st">"predictor"</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span> <span class="co"># Remove non-informative variable</span></span>
<span> <span class="fu"><a href="../reference/amr-tidymodels.html">step_mic_log2</a></span><span class="op">(</span><span class="fu"><a href="../reference/amr-tidymodels.html">all_mic_predictors</a></span><span class="op">(</span><span class="op">)</span><span class="op">)</span> <span class="co">#%&gt;% # Log2 transform all MIC predictors</span></span>
<span> <span class="co"># prep()</span></span>
<span></span>
<span><span class="va">mic_recipe</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; <span style="color: #00BBBB;">──</span> <span style="font-weight: bold;">Recipe</span> <span style="color: #00BBBB;">──────────────────────────────────────────────────────────────────────</span></span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; ── Inputs</span></span>
<span><span class="co">#&gt; Number of variables by role</span></span>
<span><span class="co">#&gt; outcome: 1</span></span>
<span><span class="co">#&gt; predictor: 17</span></span>
<span><span class="co">#&gt; undeclared role: 1</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; ── Operations</span></span>
<span><span class="co">#&gt; <span style="color: #00BBBB;">•</span> Log2 transformation of MIC columns: <span style="color: #0000BB;">all_mic_predictors()</span></span></span></code></pre></div>
<p><strong>Explanation:</strong></p>
<ul>
<li>
<code>remove_role()</code>: Removes irrelevant variables like
genus.</li>
<li>
<code><a href="../reference/amr-tidymodels.html">step_mic_log2()</a></code>: Applies
<code>log2(as.numeric(...))</code> to all MIC predictors in one go.</li>
<li>
<code>prep()</code>: Finalises the recipe based on training
data.</li>
</ul>
</div>
<div class="section level4">
<h4 id="specifying-the-model-1">2. Specifying the Model<a class="anchor" aria-label="anchor" href="#specifying-the-model-1"></a>
</h4>
<p>We use a simple logistic regression to model ESBL presence, though
recent models such as xgboost (<a href="https://parsnip.tidymodels.org/reference/details_boost_tree_xgboost.html" class="external-link">link
to <code>parsnip</code> manual</a>) could be much more precise.</p>
<div class="sourceCode" id="cb12"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Define the model</span></span>
<span><span class="va">model</span> <span class="op">&lt;-</span> <span class="fu">logistic_reg</span><span class="op">(</span>mode <span class="op">=</span> <span class="st">"classification"</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">set_engine</span><span class="op">(</span><span class="st">"glm"</span><span class="op">)</span></span>
<span></span>
<span><span class="va">model</span></span>
<span><span class="co">#&gt; Logistic Regression Model Specification (classification)</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; Computational engine: glm</span></span></code></pre></div>
<p><strong>Explanation:</strong></p>
<ul>
<li>
<code>logistic_reg()</code>: Specifies a binary classification
model.</li>
<li>
<code>set_engine("glm")</code>: Uses the base R GLM engine.</li>
</ul>
</div>
<div class="section level4">
<h4 id="building-the-workflow-1">3. Building the Workflow<a class="anchor" aria-label="anchor" href="#building-the-workflow-1"></a>
</h4>
<div class="sourceCode" id="cb13"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Create workflow</span></span>
<span><span class="va">workflow_model</span> <span class="op">&lt;-</span> <span class="fu">workflow</span><span class="op">(</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">add_recipe</span><span class="op">(</span><span class="va">mic_recipe</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">add_model</span><span class="op">(</span><span class="va">model</span><span class="op">)</span></span>
<span></span>
<span><span class="va">workflow_model</span></span>
<span><span class="co">#&gt; ══ Workflow ════════════════════════════════════════════════════════════════════</span></span>
<span><span class="co">#&gt; <span style="font-style: italic;">Preprocessor:</span> Recipe</span></span>
<span><span class="co">#&gt; <span style="font-style: italic;">Model:</span> logistic_reg()</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; ── Preprocessor ────────────────────────────────────────────────────────────────</span></span>
<span><span class="co">#&gt; 1 Recipe Step</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; • step_mic_log2()</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; ── Model ───────────────────────────────────────────────────────────────────────</span></span>
<span><span class="co">#&gt; Logistic Regression Model Specification (classification)</span></span>
<span><span class="co">#&gt; </span></span>
<span><span class="co">#&gt; Computational engine: glm</span></span></code></pre></div>
</div>
</div>
<div class="section level3">
<h3 id="training-and-evaluating-the-model-1">
<strong>Training and Evaluating the Model</strong><a class="anchor" aria-label="anchor" href="#training-and-evaluating-the-model-1"></a>
</h3>
<div class="sourceCode" id="cb14"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Fit the model</span></span>
<span><span class="va">fitted</span> <span class="op">&lt;-</span> <span class="fu">fit</span><span class="op">(</span><span class="va">workflow_model</span>, <span class="va">training_data</span><span class="op">)</span></span>
<span></span>
<span><span class="co"># Generate predictions</span></span>
<span><span class="va">predictions</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/stats/predict.html" class="external-link">predict</a></span><span class="op">(</span><span class="va">fitted</span>, <span class="va">testing_data</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/bind_cols.html" class="external-link">bind_cols</a></span><span class="op">(</span><span class="va">testing_data</span><span class="op">)</span></span>
<span></span>
<span><span class="co"># Evaluate model performance</span></span>
<span><span class="va">our_metrics</span> <span class="op">&lt;-</span> <span class="fu">metric_set</span><span class="op">(</span><span class="va">accuracy</span>, <span class="va">kap</span>, <span class="va">ppv</span>, <span class="va">npv</span><span class="op">)</span></span>
<span><span class="va">metrics</span> <span class="op">&lt;-</span> <span class="fu">our_metrics</span><span class="op">(</span><span class="va">predictions</span>, truth <span class="op">=</span> <span class="va">esbl</span>, estimate <span class="op">=</span> <span class="va">.pred_class</span><span class="op">)</span></span>
<span></span>
<span><span class="va">metrics</span></span>
<span><span class="co">#&gt; <span style="color: #949494;"># A tibble: 4 × 3</span></span></span>
<span><span class="co">#&gt; .metric .estimator .estimate</span></span>
<span><span class="co">#&gt; <span style="color: #949494; font-style: italic;">&lt;chr&gt;</span> <span style="color: #949494; font-style: italic;">&lt;chr&gt;</span> <span style="color: #949494; font-style: italic;">&lt;dbl&gt;</span></span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;">1</span> accuracy binary 0.92 </span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;">2</span> kap binary 0.840</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;">3</span> ppv binary 0.921</span></span>
<span><span class="co">#&gt; <span style="color: #BCBCBC;">4</span> npv binary 0.919</span></span></code></pre></div>
<p><strong>Explanation:</strong></p>
<ul>
<li>
<code>fit()</code>: Trains the model on the processed training
data.</li>
<li>
<code><a href="https://rdrr.io/r/stats/predict.html" class="external-link">predict()</a></code>: Produces predictions for unseen test
data.</li>
<li>
<code>metric_set()</code>: Allows evaluating multiple classification
metrics.</li>
</ul>
<p>It appears we can predict ESBL gene presence with a positive
predictive value (PPV) of 92.1% and a negative predictive value (NPV) of
91.9 using a simplistic logistic regression model.</p>
</div>
<div class="section level3">
<h3 id="visualising-predictions">
<strong>Visualising Predictions</strong><a class="anchor" aria-label="anchor" href="#visualising-predictions"></a>
</h3>
<p>We can visualise predictions by comparing predicted and actual ESBL
status.</p>
<div class="sourceCode" id="cb15"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://ggplot2.tidyverse.org" class="external-link">ggplot2</a></span><span class="op">)</span></span>
<span></span>
<span><span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggplot.html" class="external-link">ggplot</a></span><span class="op">(</span><span class="va">predictions</span>, <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/aes.html" class="external-link">aes</a></span><span class="op">(</span>x <span class="op">=</span> <span class="va">esbl</span>, fill <span class="op">=</span> <span class="va">.pred_class</span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/geom_bar.html" class="external-link">geom_bar</a></span><span class="op">(</span>position <span class="op">=</span> <span class="st">"stack"</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/labs.html" class="external-link">labs</a></span><span class="op">(</span>title <span class="op">=</span> <span class="st">"Predicted vs Actual ESBL Status"</span>,</span>
<span> x <span class="op">=</span> <span class="st">"Actual ESBL"</span>,</span>
<span> y <span class="op">=</span> <span class="st">"Count"</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggtheme.html" class="external-link">theme_minimal</a></span><span class="op">(</span><span class="op">)</span></span></code></pre></div>
<p><img src="AMR_with_tidymodels_files/figure-html/unnamed-chunk-14-1.png" width="720"></p>
</div>
<div class="section level3">
<h3 id="conclusion-1">
<strong>Conclusion</strong><a class="anchor" aria-label="anchor" href="#conclusion-1"></a>
</h3>
<p>In this example, we showcased how the new <code>AMR</code>-specific
recipe steps simplify working with <code>&lt;mic&gt;</code> columns in
<code>tidymodels</code>. The <code><a href="../reference/amr-tidymodels.html">step_mic_log2()</a></code> transformation
converts ordered MICs to log2-transformed numerics, improving
compatibility with classification models.</p>
<p>This pipeline enables realistic, reproducible, and interpretable
modelling of antimicrobial resistance data.</p>
<hr>
</div>
</div>
<div class="section level2">
<h2 id="example-3-predicting-amr-over-time">Example 3: Predicting AMR Over Time<a class="anchor" aria-label="anchor" href="#example-3-predicting-amr-over-time"></a>
</h2>
<p>In this third example, we aim to predict antimicrobial resistance
(AMR) trends over time using <code>tidymodels</code>. We will model
resistance to three antibiotics (amoxicillin <code>AMX</code>,
amoxicillin-clavulanic acid <code>AMC</code>, and ciprofloxacin
<code>CIP</code>), based on historical data grouped by year and hospital
ward.</p>
<div class="section level3">
<h3 id="objective-1">
<strong>Objective</strong><a class="anchor" aria-label="anchor" href="#objective-1"></a>
<h3 id="objective-2">
<strong>Objective</strong><a class="anchor" aria-label="anchor" href="#objective-2"></a>
</h3>
<p>Our goal is to:</p>
<ol style="list-style-type: decimal">
@ -447,12 +691,12 @@ model.</li>
</ol>
</div>
<div class="section level3">
<h3 id="data-preparation-1">
<strong>Data Preparation</strong><a class="anchor" aria-label="anchor" href="#data-preparation-1"></a>
<h3 id="data-preparation-2">
<strong>Data Preparation</strong><a class="anchor" aria-label="anchor" href="#data-preparation-2"></a>
</h3>
<p>We start by transforming the <code>example_isolates</code> dataset
into a structured time-series format.</p>
<div class="sourceCode" id="cb10"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb16"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Load required libraries</span></span>
<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://amr-for-r.org">AMR</a></span><span class="op">)</span></span>
<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://tidymodels.tidymodels.org" class="external-link">tidymodels</a></span><span class="op">)</span></span>
@ -499,15 +743,15 @@ rates by year and ward.</li>
</ul>
</div>
<div class="section level3">
<h3 id="defining-the-workflow-1">
<strong>Defining the Workflow</strong><a class="anchor" aria-label="anchor" href="#defining-the-workflow-1"></a>
<h3 id="defining-the-workflow-2">
<strong>Defining the Workflow</strong><a class="anchor" aria-label="anchor" href="#defining-the-workflow-2"></a>
</h3>
<p>We now define the modelling workflow, which consists of a
preprocessing step, a model specification, and the fitting process.</p>
<div class="section level4">
<h4 id="preprocessing-with-a-recipe-1">1. Preprocessing with a Recipe<a class="anchor" aria-label="anchor" href="#preprocessing-with-a-recipe-1"></a>
<h4 id="preprocessing-with-a-recipe-2">1. Preprocessing with a Recipe<a class="anchor" aria-label="anchor" href="#preprocessing-with-a-recipe-2"></a>
</h4>
<div class="sourceCode" id="cb11"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb17"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Define the recipe</span></span>
<span><span class="va">resistance_recipe_time</span> <span class="op">&lt;-</span> <span class="fu">recipe</span><span class="op">(</span><span class="va">res_AMX</span> <span class="op">~</span> <span class="va">year</span> <span class="op">+</span> <span class="va">gramstain</span>, data <span class="op">=</span> <span class="va">data_time</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">step_dummy</span><span class="op">(</span><span class="va">gramstain</span>, one_hot <span class="op">=</span> <span class="cn">TRUE</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span> <span class="co"># Convert categorical to numerical</span></span>
@ -540,10 +784,10 @@ variable.</li>
</ul>
</div>
<div class="section level4">
<h4 id="specifying-the-model-1">2. Specifying the Model<a class="anchor" aria-label="anchor" href="#specifying-the-model-1"></a>
<h4 id="specifying-the-model-2">2. Specifying the Model<a class="anchor" aria-label="anchor" href="#specifying-the-model-2"></a>
</h4>
<p>We use a linear regression model to predict resistance trends.</p>
<div class="sourceCode" id="cb12"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb18"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Define the linear regression model</span></span>
<span><span class="va">lm_model</span> <span class="op">&lt;-</span> <span class="fu">linear_reg</span><span class="op">(</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">set_engine</span><span class="op">(</span><span class="st">"lm"</span><span class="op">)</span> <span class="co"># Use linear regression</span></span>
@ -562,10 +806,10 @@ engine.</li>
</ul>
</div>
<div class="section level4">
<h4 id="building-the-workflow-1">3. Building the Workflow<a class="anchor" aria-label="anchor" href="#building-the-workflow-1"></a>
<h4 id="building-the-workflow-2">3. Building the Workflow<a class="anchor" aria-label="anchor" href="#building-the-workflow-2"></a>
</h4>
<p>We combine the preprocessing recipe and model into a workflow.</p>
<div class="sourceCode" id="cb13"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb19"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Create workflow</span></span>
<span><span class="va">resistance_workflow_time</span> <span class="op">&lt;-</span> <span class="fu">workflow</span><span class="op">(</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
<span> <span class="fu">add_recipe</span><span class="op">(</span><span class="va">resistance_recipe_time</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
@ -590,12 +834,12 @@ engine.</li>
</div>
</div>
<div class="section level3">
<h3 id="training-and-evaluating-the-model-1">
<strong>Training and Evaluating the Model</strong><a class="anchor" aria-label="anchor" href="#training-and-evaluating-the-model-1"></a>
<h3 id="training-and-evaluating-the-model-2">
<strong>Training and Evaluating the Model</strong><a class="anchor" aria-label="anchor" href="#training-and-evaluating-the-model-2"></a>
</h3>
<p>We split the data into training and testing sets, fit the model, and
evaluate performance.</p>
<div class="sourceCode" id="cb14"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb20"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="co"># Split the data</span></span>
<span><span class="fu"><a href="https://rdrr.io/r/base/Random.html" class="external-link">set.seed</a></span><span class="op">(</span><span class="fl">123</span><span class="op">)</span></span>
<span><span class="va">data_split_time</span> <span class="op">&lt;-</span> <span class="fu">initial_split</span><span class="op">(</span><span class="va">data_time</span>, prop <span class="op">=</span> <span class="fl">0.8</span><span class="op">)</span></span>
@ -636,11 +880,11 @@ sets.</li>
</ul>
</div>
<div class="section level3">
<h3 id="visualising-predictions">
<strong>Visualising Predictions</strong><a class="anchor" aria-label="anchor" href="#visualising-predictions"></a>
<h3 id="visualising-predictions-1">
<strong>Visualising Predictions</strong><a class="anchor" aria-label="anchor" href="#visualising-predictions-1"></a>
</h3>
<p>We plot resistance trends over time for amoxicillin.</p>
<div class="sourceCode" id="cb15"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb21"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://ggplot2.tidyverse.org" class="external-link">ggplot2</a></span><span class="op">)</span></span>
<span></span>
<span><span class="co"># Plot actual vs predicted resistance over time</span></span>
@ -651,10 +895,10 @@ sets.</li>
<span> x <span class="op">=</span> <span class="st">"Year"</span>,</span>
<span> y <span class="op">=</span> <span class="st">"Resistance Proportion"</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggtheme.html" class="external-link">theme_minimal</a></span><span class="op">(</span><span class="op">)</span></span></code></pre></div>
<p><img src="AMR_with_tidymodels_files/figure-html/unnamed-chunk-14-1.png" width="720"></p>
<p><img src="AMR_with_tidymodels_files/figure-html/unnamed-chunk-20-1.png" width="720"></p>
<p>Additionally, we can visualise resistance trends in
<code>ggplot2</code> and directly add linear models there:</p>
<div class="sourceCode" id="cb16"><pre class="downlit sourceCode r">
<div class="sourceCode" id="cb22"><pre class="downlit sourceCode r">
<code class="sourceCode R"><span><span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggplot.html" class="external-link">ggplot</a></span><span class="op">(</span><span class="va">data_time</span>, <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/aes.html" class="external-link">aes</a></span><span class="op">(</span>x <span class="op">=</span> <span class="va">year</span>, y <span class="op">=</span> <span class="va">res_AMX</span>, color <span class="op">=</span> <span class="va">gramstain</span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/geom_path.html" class="external-link">geom_line</a></span><span class="op">(</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/labs.html" class="external-link">labs</a></span><span class="op">(</span>title <span class="op">=</span> <span class="st">"AMX Resistance Trends"</span>,</span>
@ -665,11 +909,11 @@ sets.</li>
<span> formula <span class="op">=</span> <span class="va">y</span> <span class="op">~</span> <span class="va">x</span>,</span>
<span> alpha <span class="op">=</span> <span class="fl">0.25</span><span class="op">)</span> <span class="op">+</span></span>
<span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggtheme.html" class="external-link">theme_minimal</a></span><span class="op">(</span><span class="op">)</span></span></code></pre></div>
<p><img src="AMR_with_tidymodels_files/figure-html/unnamed-chunk-15-1.png" width="720"></p>
<p><img src="AMR_with_tidymodels_files/figure-html/unnamed-chunk-21-1.png" width="720"></p>
</div>
<div class="section level3">
<h3 id="conclusion-1">
<strong>Conclusion</strong><a class="anchor" aria-label="anchor" href="#conclusion-1"></a>
<h3 id="conclusion-2">
<strong>Conclusion</strong><a class="anchor" aria-label="anchor" href="#conclusion-2"></a>
</h3>
<p>In this example, we demonstrated how to analyze AMR trends over time
using <code>tidymodels</code>. By aggregating resistance rates by year