Update README.md
Browse files
README.md
CHANGED
|
@@ -26,7 +26,7 @@ size_categories:
|
|
| 26 |
|
| 27 |
**Structured, issue-level records extracted from U.S. Board of Veterans' Appeals (BVA) decisions** — each decision parsed into its issues, conditions, outcomes, citations, and reasoning, with per-document provenance and completeness flags. Built for training and evaluating legal-AI models on veterans' disability adjudication.
|
| 28 |
|
| 29 |
-
This is a **
|
| 30 |
|
| 31 |
> **Honesty note (please read).** Labels are **silver** (engine-produced, benchmarked against an LLM-labeled reference at ~96% outcome accuracy), **not** human-certified gold. Provenance and completeness ship with every row so you can verify and filter. See **Quality & accuracy**.
|
| 32 |
|
|
@@ -36,17 +36,17 @@ This is a **5600-decision sample**, balanced across **seven years (2019–2025,
|
|
| 36 |
|
| 37 |
| | |
|
| 38 |
|---|---|
|
| 39 |
-
| **Decisions** |
|
| 40 |
-
| **Years** | 2019–2025 (
|
| 41 |
| **Format** | CSV (one row per decision) — 23 columns |
|
| 42 |
| **Coverage** | 99% of rows carry extracted issues and citations; 96% carry reasoning atoms |
|
| 43 |
| **Source** | Public BVA decisions published on va.gov |
|
| 44 |
|
| 45 |
### Outcome distribution (decision-level)
|
| 46 |
-
`remanded`
|
| 47 |
|
| 48 |
### Appeals regime
|
| 49 |
-
`legacy`
|
| 50 |
|
| 51 |
### Conditions
|
| 52 |
**107 distinct medical conditions.** Top: back disability (490), knee disability (438), hearing loss (341), arthritis (302), psychiatric disorder (289), PTSD (287), peripheral neuropathy (272), sleep apnea (212), hypertension (195), foot disability (177).
|
|
@@ -116,5 +116,5 @@ Treat the accuracy figure as a **silver benchmark**, and verify against `source_
|
|
| 116 |
---
|
| 117 |
|
| 118 |
## Changelog
|
| 119 |
-
- **v1** —
|
| 120 |
---
|
|
|
|
| 26 |
|
| 27 |
**Structured, issue-level records extracted from U.S. Board of Veterans' Appeals (BVA) decisions** — each decision parsed into its issues, conditions, outcomes, citations, and reasoning, with per-document provenance and completeness flags. Built for training and evaluating legal-AI models on veterans' disability adjudication.
|
| 28 |
|
| 29 |
+
This is a **2900-decision sample**, balanced across **seven years (2019–2025, ~800 decisions/year)**, so it's representative of the full corpus rather than skewed to one year. A larger full-corpus release and a commercial license are available (see **Access & licensing** below).
|
| 30 |
|
| 31 |
> **Honesty note (please read).** Labels are **silver** (engine-produced, benchmarked against an LLM-labeled reference at ~96% outcome accuracy), **not** human-certified gold. Provenance and completeness ship with every row so you can verify and filter. See **Quality & accuracy**.
|
| 32 |
|
|
|
|
| 36 |
|
| 37 |
| | |
|
| 38 |
|---|---|
|
| 39 |
+
| **Decisions** | 2900
|
| 40 |
+
| **Years** | 2019–2025 (balanced) |
|
| 41 |
| **Format** | CSV (one row per decision) — 23 columns |
|
| 42 |
| **Coverage** | 99% of rows carry extracted issues and citations; 96% carry reasoning atoms |
|
| 43 |
| **Source** | Public BVA decisions published on va.gov |
|
| 44 |
|
| 45 |
### Outcome distribution (decision-level)
|
| 46 |
+
`remanded` · `denied` · `granted` · `dismissed` · `reopened` · `withdrawn`
|
| 47 |
|
| 48 |
### Appeals regime
|
| 49 |
+
`legacy` · `AMA` · `unknown` (tagged from the decision text, not the year).
|
| 50 |
|
| 51 |
### Conditions
|
| 52 |
**107 distinct medical conditions.** Top: back disability (490), knee disability (438), hearing loss (341), arthritis (302), psychiatric disorder (289), PTSD (287), peripheral neuropathy (272), sleep apnea (212), hypertension (195), foot disability (177).
|
|
|
|
| 116 |
---
|
| 117 |
|
| 118 |
## Changelog
|
| 119 |
+
- **v1** — 2900 decisions, balanced 2019–2025; 23-column document schema; silver labels (~96% benchmark).
|
| 120 |
---
|