{
  "schema_version": "1.0.0",
  "exported_at": "2026-03-05T22:01:19.677375+00:00",
  "data": [
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "accuracy",
      "value": 0.9997593509350935
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "precision",
      "value": 1.0
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "recall",
      "value": 0.9991956796506951
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "f1",
      "value": 0.9995976780274728
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "specificity",
      "value": 1.0
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "balanced_accuracy",
      "value": 0.9995978398253476
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "mcc",
      "value": 0.9994261773037189
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "fpr",
      "value": 0.0
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "fnr",
      "value": 0.0008043203493048374
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "n_samples",
      "value": 29088
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "auc",
      "value": 0.9999977791638275
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "kappa",
      "value": 0.9994260126675026
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "tpr",
      "value": 0.9991956796506951
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "tnr",
      "value": 1.0
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "detection_rate",
      "value": 0.9991956796506951
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "miss_rate",
      "value": 0.0008043203493048374
    },
    {
      "model": "gat",
      "scenario": "val",
      "metric_name": "pr_auc",
      "value": 0.9999949843911157
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "accuracy",
      "value": 0.7375893839383938
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "precision",
      "value": 0.538920413211116
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "recall",
      "value": 0.8512007353786051
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "f1",
      "value": 0.6599848545592232
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "specificity",
      "value": 0.6890851116016679
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "balanced_accuracy",
      "value": 0.7701429234901365
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "mcc",
      "value": 0.495546350204486
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "fpr",
      "value": 0.3109148883983321
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "fnr",
      "value": 0.14879926462139492
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "n_samples",
      "value": 29088
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "auc",
      "value": 0.864026205190438
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "optimal_threshold",
      "value": 0.03546275198459625
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "youden_j",
      "value": 0.5404007498873165
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "kappa",
      "value": 0.4633535089228278
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "tpr",
      "value": 0.8512007353786051
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "tnr",
      "value": 0.6890851116016679
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "detection_rate",
      "value": 0.8512007353786051
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "miss_rate",
      "value": 0.14879926462139492
    },
    {
      "model": "vgae",
      "scenario": "val",
      "metric_name": "pr_auc",
      "value": 0.7561650999976426
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "accuracy",
      "value": 0.9997593509350935
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "precision",
      "value": 1.0
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "recall",
      "value": 0.9991956796506951
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "f1",
      "value": 0.9995976780274728
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "specificity",
      "value": 1.0
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "balanced_accuracy",
      "value": 0.9995978398253476
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "mcc",
      "value": 0.9994261773037189
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "fpr",
      "value": 0.0
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "fnr",
      "value": 0.0008043203493048374
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "n_samples",
      "value": 29088
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "auc",
      "value": 0.9996878118735314
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "kappa",
      "value": 0.9994260126675026
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "tpr",
      "value": 0.9991956796506951
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "tnr",
      "value": 1.0
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "detection_rate",
      "value": 0.9991956796506951
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "miss_rate",
      "value": 0.0008043203493048374
    },
    {
      "model": "fusion",
      "scenario": "val",
      "metric_name": "pr_auc",
      "value": 0.9996734070453062
    }
  ]
}