data/artifacts/models/risk_model/metrics.json

{
  "model_name": "risk_model",
  "model_version": "risk_model_v2_baseline_001",
  "dataset_version": "2.0.0",
  "trained_at": "2026-05-21T05:59:12.289643+00:00",
  "seed": 42,
  "split_counts": {
    "train": 1133,
    "validation": 243,
    "test": 244
  },
  "metrics": {
    "validation": {
      "recall_positive": 0.8333,
      "precision_positive": 0.9459,
      "f1_positive": 0.8861,
      "roc_auc": 0.99,
      "per_class": {
        "not_at_risk": {
          "precision": 0.966,
          "recall": 0.99,
          "f1": 0.9779,
          "support": 201
        },
        "at_risk": {
          "precision": 0.9459,
          "recall": 0.8333,
          "f1": 0.8861,
          "support": 42
        }
      },
      "confusion_matrix": [
        [
          199,
          2
        ],
        [
          7,
          35
        ]
      ],
      "risk_level_recall": {
        "high": {
          "recall": 0.8158,
          "support": 38
        },
        "critical": {
          "recall": 1.0,
          "support": 4
        }
      }
    },
    "test": {
      "recall_positive": 0.7812,
      "precision_positive": 0.9259,
      "f1_positive": 0.8475,
      "roc_auc": 0.9899,
      "per_class": {
        "not_at_risk": {
          "precision": 0.9677,
          "recall": 0.9906,
          "f1": 0.979,
          "support": 212
        },
        "at_risk": {
          "precision": 0.9259,
          "recall": 0.7812,
          "f1": 0.8475,
          "support": 32
        }
      },
      "confusion_matrix": [
        [
          210,
          2
        ],
        [
          7,
          25
        ]
      ],
      "risk_level_recall": {
        "high": {
          "recall": 0.7407,
          "support": 27
        },
        "critical": {
          "recall": 1.0,
          "support": 5
        }
      }
    }
  },
  "limitations": [
    "Trained on synthetic data only.",
    "Binary risk_label derived from synthetic risk_score thresholds.",
    "All features are numeric; no text or contextual features used.",
    "Class imbalance (~16% positive) addressed via class_weight='balanced'.",
    "Critical class (~2%) recall should be monitored separately."
  ]
}