The process involves these four steps:
- Selecting relevant data that can be interpreted through machine learning.
- Defining the type of annotation
- Labeling different parts by Keyphrasing, classification, and language identification.
- The last and crucial step is cross-checking and reviewing the extracted data.