docs(accuracy): publish phase 4 benchmark evidence

2026-08-02 05:10:05 +02:00
parent 70fb4b94e5
commit f41392a415
34 changed files with 25672 additions and 8 deletions
@@ -0,0 +1,48 @@
+{
+  "claim_boundary": "Synthetic deterministic reference cases validate evaluator behavior, not production model accuracy.",
+  "observed_codes": [
+    {
+      "error_code": "D-VALIDATION-FN",
+      "kind": "validation_false_negative"
+    },
+    {
+      "error_code": "D-VALIDATION-FP",
+      "kind": "validation_false_positive"
+    },
+    {
+      "error_code": "M-AREA-BIAS",
+      "kind": "area_bias"
+    },
+    {
+      "error_code": "M-CLASS",
+      "kind": "raster_misclassification"
+    },
+    {
+      "error_code": "M-FN-MISSED",
+      "kind": "event_false_negative"
+    },
+    {
+      "error_code": "M-FN-MISSED",
+      "kind": "false_negative"
+    },
+    {
+      "error_code": "M-FP-CONFUSER",
+      "kind": "event_false_positive"
+    },
+    {
+      "error_code": "M-FP-CONFUSER",
+      "kind": "false_positive"
+    },
+    {
+      "error_code": "M-MISCALIBRATED",
+      "kind": "miscalibrated"
+    },
+    {
+      "error_code": "P-PARTIAL",
+      "kind": "terrain_missing"
+    }
+  ],
+  "observed_failure_count": 14,
+  "schema_version": 2,
+  "source": "docs/accuracy-program/05-metric-framework.md section 4"
+}