1. The initial or primary collection of data before any processing, transformation, or modification has been applied
The original dataset contained 10,000 records before we removed duplicates.
O conjunto de dados original continha 10.000 registros antes de removermos os duplicados.
2. In machine learning and data science, the unmodified source data used as the foundation for analysis or model training
We must preserve the original dataset for reproducibility and validation purposes.
Devemos preservar o conjunto de dados original para fins de reprodutibilidade e validação.