{
  "context": "Evaluates how well the solution uses scikit-learn's incremental learning tools to implement the streaming classifier. Focuses on correct use of partial_fit for batch updates and warm_start for continuing training without resetting weights, plus standard prediction/scoring APIs.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Partial_fit estimator",
      "description": "Uses a scikit-learn classifier that implements partial_fit (e.g., SGDClassifier, PassiveAggressiveClassifier, Perceptron) rather than a batch-only estimator.",
      "max_score": 25
    },
    {
      "name": "Classes registration",
      "description": "Supplies the full set of target labels via the classes parameter on the first partial_fit call to register label support for subsequent batches.",
      "max_score": 15
    },
    {
      "name": "Streaming updates",
      "description": "Processes each incoming batch by calling estimator.partial_fit on the same instance (no re-instantiation) so weights accumulate across batches.",
      "max_score": 20
    },
    {
      "name": "Warm-start resume",
      "description": "Enables warm_start=True (or equivalent) and performs repeat fit calls for extra epochs so resumed training continues from existing coefficients instead of reinitializing.",
      "max_score": 20
    },
    {
      "name": "Predict & score",
      "description": "Generates outputs via estimator.predict and computes accuracy using estimator.score or sklearn.metrics.accuracy_score to validate performance.",
      "max_score": 20
    }
  ]
}