Update README.md

README.md

@@ -1,3 +1,100 @@
----
-license: mit
----
+---
+language:
+  - en
+license: mit
+pretty_name: FFR Center Failure Surface Routing v0.1
+dataset_name: ffr-center-failure-surface-routing-v0.1
+tags:
+  - clarusc64
+  - clinical
+  - cardiology
+  - ffr
+  - site-performance
+  - routing
+  - protocol-drift
+  - vendor-audit
+task_categories:
+  - tabular-classification
+  - tabular-regression
+size_categories:
+  - n<1K
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: data/train.csv
+      - split: test
+        path: data/test.csv
+---
+
+Goal
+
+When a center drifts  
+route the safest operational response.
+
+This dataset treats each site  
+as a failure surface.
+
+Not a single bad scan.
+
+Inputs
+
+Rolling site window metrics:
+
+- image quality proxies
+- protocol signature
+- rolling performance (AUC, MAE)
+- calibration error shift
+- plausibility conflict rate
+- drift type and risk
+
+Required outputs
+
+- failure_surface_type
+- subgroup_affected
+- intervention_route
+- recalibration_needed
+- protocol_reset_plan
+- escalation_priority
+- confidence_score
+
+Failure surfaces
+
+Examples:
+
+- minor shift surface
+- protocol drift surface
+- silent decay surface
+- quality collapse surface
+- outside envelope surface
+
+Intervention routes
+
+Examples:
+
+- continue
+- tighten QA and monitor
+- restrict use to low risk cases
+- rescan or block output
+- disable model at site
+
+Why it matters
+
+The same model can be safe at one site  
+and unsafe at another.
+
+Hospitals need a clear route.  
+Vendors need early containment.
+
+This dataset makes those actions  
+testable and repeatable.
+
+Evaluation
+
+The scorer checks the response includes:
+
+- surface label
+- subgroup label
+- route + reset plan
+- escalation priority
+- confidence score 0 to 1