data drift
[/ˈdeɪtə drɪft/]
nounpl: data drifts
desvio de dados
1. A phenomenon in machine learning and data science where the statistical properties of input data change over time, causing a trained model's performance to degrade
The recommendation system experienced data drift after the company changed its product catalog, resulting in less accurate predictions.
O sistema de recomendação experimentou desvio de dados após a empresa alterar seu catálogo de produtos, resultando em previsões menos precisas.
2. The gradual shift in the distribution of data features that differs from the training dataset, requiring model retraining or adjustment
Monitoring for data drift is essential to maintain model reliability in production environments.
Monitorar o desvio de dados é essencial para manter a confiabilidade do modelo em ambientes de produção.
This is a specialized technical term primarily used in data science, machine learning, and artificial intelligence communities in both Brazil and the United States. It has become increasingly important as companies rely more on ML models in production. The term is used consistently across both English and Portuguese-speaking tech communities without significant regional variation.
Look up more words on Fala2Me
The free English-Portuguese dictionary with real Brazilian accents, NYC slang, conjugator and more
Open Fala2Me →