Skip to main content
Glite Dictionary
📊 Too close fit (data) 👟 Too tight fit
overfitting

overfitting

8.1
A model that learns the training data too well and fails on new data
  • noun
  • /ˌoʊvərˈfɪtɪŋ/
  • Specialized
translation icon : sobreajuste
  • OK, So this the idea of overfitting and stability that came out in this discussion are key.

So stability or differential privacy will preserve pains, privacy and protect against overfitting of the data.

Examples

  • This is "justified" in the name of avoiding overfitting, but this is actually an extra fitting step based on RE statistics.

    Blog text (21)

  • This is very, very serious overfitting of data.

  • This observation suggested that error increased and overfitting occurred when more items were deleted.

    Academic text (2005)

  • I actually kind of lean into overfitting because think of it as there are ideas that have been so rehearsed.

  • In machine learning, for example, maximizing the likelihood is well known to cause overfitting and result in poor predictive performance.

    Academic text (2017)

  • Consequently, the exploratory findings, in particular, may be subject to overfitting of parameters and unstable conclusions.

    Academic text (1990)

  • When included in the model, the presence of cavitation on chest radiograph increased the ORs for all three risk factors; however, this variable was not included in the final model because of concerns about overfitting and loss of precision.

    Academic text (1999)

  • Srivastava et al. (43) suggested a simple regularization technique named "Dropout" to solve the overfitting problem and to improve the performance of the NN.

    Academic text (2018)

  • This type of overfitting permits experts to use complicated calculations as a smoke screen to obscure the actual process behind arriving at their final figure—a figure that is actually just the result of statistical noise.

    Academic text (2017)

  • Overfitting occurs when a model learns the training data too well, leading to poor results on new overfitting information.

Synonyms

overtraining
vsoverfitting
  • Specialized
6.6

Teaching a computer program too much with the same examples that it fails on new data

is the model fitting training examples too closely harming general performance

Antonyms

generalization
  • Specialized
1 5.7

A model's ability to do well on new, unseen examples

How Memorized

overfitting
  • Specialized
8.1
memorization
  • Specialized
6.1
generalization
  • Specialized
1 6.3

Surface Forms

overfitting singular
overfittings plural

Morphology

overfitting = overfit (transparent) = over + fit + ing

Formed by adding the verbal suffix '-ing' to 'overfit' (over- + fit), a standard and directly interpretable process meaning 'excessive fitting'.

Etymology

Overfitting combines over- 'too much' and fit 'to match'. Imagine a model that 'fits' its training examples so exactly it learns every tiny detail, like a suit made for one person that will not fit others, so in machine learning overfitting means the model matches the training data too closely and fails on new data.