Update README.md

README.md

@@ -8,14 +8,17 @@ widget:
   example_title: "example 1"
 ---
 
-This is a model to detect treatment and disease mentions
+This is a model to detect treatment and disease mentions in texts from health domains. The dataset used for training consists of PubMed's abstracts and tweets with disease mentions. It has been semi-automatically labeled using a system of regex rules for matching treatment and disease mentions, for example
 
-## Model performances (metric: seqeval)
+`<treatment> (good|great) (for|to|in) <disease>`
+`<treatment> ( \S+){0,4} improve( \S+){0,4} <disease>`
+`<disease> requiring <treatment>`
+
+and so on.
 
-Overall
 
 precision|recall|f1
 -|-|-
 0.92|0.95|0.93
 
-Work in progress!
+