{
  "slug": "temporal-validation",
  "term": "Temporal validation",
  "aliases": [
    "time-forward validation",
    "chronological validation"
  ],
  "definition": "Temporal validation evaluates a predictor on later observations using training information available before the evaluation boundary.",
  "note": "Training on earlier observations and testing on later ones approximates a future-prediction task; a random split can expose training to future conditions. [sklearn-validation] For disk errors, Xu et al. report a setting-specific gap between random and time-forward evaluation, not a universal correction factor. [xu-atc18] Chronological separation is different from grouping by drive identity: it does not by itself test generalization to unseen drives. Feature selection and preprocessing must be fitted using training data only. [sklearn-pitfalls]",
  "limitations": [
    {
      "text": "Storage evaluation should state prediction time, outcome horizon, label availability, drive overlap and test prevalence; a chronological split alone is not proof that all leakage is absent.",
      "citationIds": [
        "xu-atc18",
        "sklearn-validation",
        "sklearn-pitfalls"
      ]
    }
  ],
  "related": [
    "smart",
    "class-prevalence",
    "precision",
    "recall",
    "f1-score",
    "drive-day",
    "prediction-horizon",
    "grouped-cross-validation",
    "failure-label"
  ],
  "citations": [
    {
      "id": "sklearn-validation",
      "title": "scikit-learn: cross-validation, grouped data and time-series evaluation.",
      "url": "https://scikit-learn.org/stable/modules/cross_validation.html",
      "accessed": "2026-10-14"
    },
    {
      "id": "xu-atc18",
      "title": "Xu et al. Improving Service Availability of Cloud Systems by Predicting Disk Error. USENIX ATC 2018, sections 2.2 and 4.4.",
      "url": "https://www.usenix.org/system/files/conference/atc18/atc18-xu-yong.pdf",
      "accessed": "2026-10-14"
    },
    {
      "id": "sklearn-pitfalls",
      "title": "scikit-learn: common pitfalls, preprocessing and data leakage.",
      "url": "https://scikit-learn.org/stable/common_pitfalls.html#data-leakage",
      "accessed": "2026-10-14"
    }
  ],
  "status": "shipping",
  "advantages": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-10-14"
}