double dipping
- noun
- Specialized
:
doble uso de datos
- Reviewers noted the paper suffered from double dipping, using the same dataset for training and testing.
Examples
-
Our lab fixed the issue by avoiding double dipping and adopting proper cross-validation procedures.
-
Because of double dipping, their model reported unrealistically high accuracy on the evaluation set.
-
Avoid double dipping when training your model.
-
The study showed that double dipping can lead to inaccurate results.
-
The researchers discovered that their analysis was flawed due to double dipping, as they evaluated the model on the same data used for training.
-
To ensure the validity of the results, the team avoided double dipping and utilized separate datasets for training and testing.
-
The enhanced algorithm outperformed others because it effectively addressed double dipping in its evaluation metrics.
Surface Forms
Morphology
The constituents 'double' (twice) and 'dip' (to take/use) transparently suggest the general idea of taking or using something twice, so a learner could guess a sense of 'using twice'. However, the specific technical meaning in statistics/data science—reusing the same data for both training and evaluation causing biased estimates—is domain-specific and not fully predictable from the parts alone, so it is only partially transparent.
Etymology
Double dipping gives the image of taking a second dip into the same set of questions or data, so you get the same thing twice. In data work, that image explains why double dipping means using the same data to 'teach' a computer program and to 'check' it, which makes results look too good.
