Update README.md

README.md

@@ -17,7 +17,7 @@ size_categories:
 
 <!-- Provide a quick summary of the dataset. -->
 
-This parallel corpus dataset contains about 21k rows of parallel English and Spanish texts.
+This parallel corpus dataset contains about 21k rows of parallel English and Spanish texts obtained by crawling different websites. It has been filtered strictly.
 
 ## Dataset Details
 
@@ -25,9 +25,14 @@ This parallel corpus dataset contains about 21k rows of parallel English and Spa
 
 <!-- Provide a longer summary of what this dataset is. -->
 
+This is a parallel corpus of bilingual texts crawled from multilingual websites, which contains 21, 005 TUs. A strict validation process has been followed, which resulted in discarding:
 
+  - TUs from crawled websites that do not comply to the PSI directive,
+  - TUs with more than 99% of mispelled tokens,
+  - TUs identified during the manual validation process and all the TUs from websites which error rate in the sample extracted for manual validation is strictly above the following thresholds: 50% of TUs with language identification errors, 50% of TUs with alignment errors, 50% of TUs with tokenization errors, 20% of TUs identified as machine translated content, 50% of TUs with translation errors.
 
-- **Curated by:** Directorate-General for Communications Networks, Content and Technology
+- **Period of crawling:** 15/11/2016 - 23/01/2017 (DD/MM/YY).
+- **Curated by:** Directorate-General for Communications Networks, Content and Technology.
 - **Funded by [optional]:** [More Information Needed]
 - **Shared by [optional]:** [More Information Needed]
 - **Language(s) (NLP):** English & Spanish