Update README.md

README.md

@@ -36,7 +36,7 @@ prediction  = output[0]['label'] # 0=”neutral”; 1=”positive”; 2=”negat
 ### Training data and procedure
 RePublic was domain-adapted on 91 661 Flemish news articles from three popular Flemish news providers between 2000 and 2020 (“Het Laatste Nieuws”, “Het Nieuwsblad” and “De Morgen”). These articles mention at least one out of a pre-defined list of 24 public service organizations, which contains, a.o., De Lijn (public transport organization), VDAB (Flemish job placement service), and Agentschap Zorg en Gezondheid (healthcare service). The domain adaptation was achieved by performing BERT’s language modeling tasks (masked language modeling & next sentence prediction).
 
-The model was then fine-tuned on a sentiment classification task (“positive”, “negative”, “neutral”). The supervised data consisted of 4404 annotated sentences mentioning Flemish public agencies of which 1257 sentences were positive, 1485 sentences were negative and 1662 sentences were neutral. Fine-tuning was performed for 4 epochs using a batch size of 8 and a learning rate of 5e-5. 
+The model was then fine-tuned on a sentiment classification task (“positive”, “negative”, “neutral”). The supervised data consisted of 4404 annotated sentences mentioning Flemish public agencies of which 1257 sentences were positive, 1485 sentences were negative and 1662 sentences were neutral. Fine-tuning was performed for 4 epochs using a batch size of 8 and a learning rate of 5e-5. In order to evaluate the model, a 10-fold cross validation experiment was conducted. The results of this experiment can be found below.
 
 | **Class** | **Precision (%)** | **Recall (%)** | **F1-score (%)** |
 |:---:|:---:|:---:|:---:|