From f3286cd3e8046a1fca902e11ee162fce50114ed1 Mon Sep 17 00:00:00 2001
From: Konstantin Schulz <schulzkx@hu-berlin.de>
Date: Wed, 21 Aug 2024 11:15:02 +0200
Subject: [PATCH] add NER evaluation results

---
 README.md | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

diff --git a/README.md b/README.md
index 7812ed8..272596d 100644
--- a/README.md
+++ b/README.md
@@ -1,9 +1,25 @@
 # SEFLAG
+
 **S**ystematic **E**valuation **F**ramework for NLP models and datasets in **L**atin and **A**ncient **G**reek
 
 ## Evaluation Results
+
 ### Lemmatization on UD test data
-#### Ancient Greek: greCy 
+
+#### Ancient Greek: greCy
+
 {'accuracy': 0.8942049121548943}
+
 #### Latin: LatinCy
+
 {'accuracy': 0.8843245653143111}
+
+### NER with average metrics (weighted -- micro -- macro)
+
+|               |                    |                |
+|---------------|--------------------|----------------|
+| **Model**     | flair_grc_bert_ner | LatinCy        |
+| **Dataset**   | Yousef et al.      | Herodotos      |
+| **Precision** | 96 -- 94 -- 66     | 96 -- 96 -- 59 |
+| **Recall**    | 94 -- 94 -- 76     | 96 -- 96 -- 60 |
+| **F1**        | 95 -- 94 -- 64     | 96 -- 96 -- 58 |