LeakSplits-class

LeakFit-class

LeakAudit-class

These classes capture splits, model fits, and audit diagnostics produced by
<code>make_split_plan()</code>, <code>fit_resample()</code>, and <code>audit_leakage()</code>.

Prevents and detects information leakage in biomedical machine learning.
Provides leakage-resistant split policies (subject-grouped, batch-blocked, study leave-out, time-ordered),
guarded preprocessing (train-only imputation, normalization, filtering, feature selection),
cross-validated fitting with common learners, permutation-gap auditing, batch and fold association tests,
and duplicate detection.

Selcuk Korkmaz

bioLeak

Leakage-Safe Modeling and Auditing for Genomic and Clinical Data

LeakSplits-class function

<dl>
<dt><code>mode</code></dt>
<dd>Splitting mode (e.g., "grouped_cv", "batch_blocked")</dd></dl>
<dt><code>indices</code></dt>
<dd>List of resampling descriptors (train/test indices when available)</dd>
<dt><code>info</code></dt>
<dd>Metadata associated with split or fit</dd>
<dt><code>splits</code></dt>
<dd>A [`LeakSplits`] object used for resampling</dd>
<dt><code>metrics</code></dt>
<dd>Model performance metrics per resample</dd>
<dt><code>metric_summary</code></dt>
<dd>Summary of metrics across resamples</dd>
<dt><code>audit</code></dt>
<dd>Audit information per resample</dd>
<dt><code>predictions</code></dt>
<dd>List of prediction objects</dd>
<dt><code>preprocess</code></dt>
<dd>Preprocessing steps used during fitting</dd>
<dt><code>learners</code></dt>
<dd>Learner definitions used in the pipeline</dd>
<dt><code>outcome</code></dt>
<dd>Outcome variable name</dd>
<dt><code>task</code></dt>
<dd>Modeling task name</dd>
<dt><code>feature_names</code></dt>
<dd>Feature names included in the model</dd>
<dt><code>info</code></dt>
<dd>Additional metadata about the fit</dd>
<dt><code>fit</code></dt>
<dd>A [`LeakFit`] object used to generate the audit</dd>
<dt><code>permutation_gap</code></dt>
<dd>Data frame summarising permutation gaps</dd>
<dt><code>perm_values</code></dt>
<dd>Numeric vector of permutation-based scores</dd>
<dt><code>batch_assoc</code></dt>
<dd>Data frame of batch associations</dd>
<dt><code>target_assoc</code></dt>
<dd>Data frame of feature-wise outcome associations</dd>
<dt><code>duplicates</code></dt>
<dd>Data frame detailing duplicate records</dd>
<dt><code>trail</code></dt>
<dd>List capturing audit trail information</dd>

Slots

S4 Classes for bioLeak Pipeline — LeakSplits-class

<dl>
<dt><code>mode</code></dt>
<dd>Splitting mode (e.g., "grouped_cv", "batch_blocked")</dd>


<dt><code>indices</code></dt>
<dd>List of resampling descriptors (train/test indices when available)</dd>


<dt><code>info</code></dt>
<dd>Metadata associated with split or fit</dd>


<dt><code>splits</code></dt>
<dd>A [`LeakSplits`] object used for resampling</dd>


<dt><code>metrics</code></dt>
<dd>Model performance metrics per resample</dd>


<dt><code>metric_summary</code></dt>
<dd>Summary of metrics across resamples</dd>


<dt><code>audit</code></dt>
<dd>Audit information per resample</dd>


<dt><code>predictions</code></dt>
<dd>List of prediction objects</dd>


<dt><code>preprocess</code></dt>
<dd>Preprocessing steps used during fitting</dd>


<dt><code>learners</code></dt>
<dd>Learner definitions used in the pipeline</dd>


<dt><code>outcome</code></dt>
<dd>Outcome variable name</dd>


<dt><code>task</code></dt>
<dd>Modeling task name</dd>


<dt><code>feature_names</code></dt>
<dd>Feature names included in the model</dd>


<dt><code>info</code></dt>
<dd>Additional metadata about the fit</dd>


<dt><code>fit</code></dt>
<dd>A [`LeakFit`] object used to generate the audit</dd>


<dt><code>permutation_gap</code></dt>
<dd>Data frame summarising permutation gaps</dd>


<dt><code>perm_values</code></dt>
<dd>Numeric vector of permutation-based scores</dd>


<dt><code>batch_assoc</code></dt>
<dd>Data frame of batch associations</dd>


<dt><code>target_assoc</code></dt>
<dd>Data frame of feature-wise outcome associations</dd>


<dt><code>duplicates</code></dt>
<dd>Data frame detailing duplicate records</dd>


<dt><code>trail</code></dt>
<dd>List capturing audit trail information</dd>


</dl>

LeakSplits-class: S4 Classes for bioLeak Pipeline

Description

Arguments

Value

Slots

See Also