https://sergiomorapardo.github.io/tags/scikit-learn/Recent content in Scikit-Learn on Sergio A. Mora Pardo | AI/ML Technical LeadHugoen-USSun, 06 Sep 2026 00:00:00 -0500https://sergiomorapardo.github.io/writing/cross-validation-that-matches-the-data/Sun, 06 Sep 2026 00:00:00 -0500https://sergiomorapardo.github.io/writing/cross-validation-that-matches-the-data/<p>Cross-validation is often introduced as a setting: choose five folds and call <code>cross_val_score</code>. The more important decision happens earlier. A validation design is a claim about which observations may legitimately stand in for the future.</p> <p>The source notebook builds that claim progressively, from repeated hold-out experiments to grouped, temporal, stratified, and nested validation. This note condenses the workflow into the decisions that make an evaluation credible.</p> <h2 id="start-with-the-source-of-dependence">Start with the source of dependence</h2> <p>A random split assumes observations can be exchanged without changing the problem. That assumption breaks when examples share a patient, customer, device, document, or time period. The split strategy must preserve the boundary that will exist when the model is used.</p>