4 Steps & 4 Consideration

In the course of ERADing, we follow and consider these items:

Step 1: Selecting raw naturally-occurring language representation sources

Step 2: Filtering and transcribing

Step 3: Normalizing and pre-processing

Step 4: Human translation into English (for parallel corpora only)

Consideration 1: Opting for a high variety and avoiding sub-dialectal or close varieties

Consideration 2: Minimal inclusion of French lexicons

Consideration 3: Choosing MSA orthography rules for scripting

Consideration 4: Verifying and extending datasets' scales