Skip to main content
whitening

whitening

8.2
A data processing step that removes links between parts of the data and changes their values to the same scale
  • noun
  • /ˈhwaɪtənɪŋ/
  • Specialized
translation icon : blanqueamiento (estadística)
  • To enhance the accuracy of our analysis, we applied whitening to the data set before training the algorithm.

Examples

  • vs = var(As T); Bs = As * diag(vs - 1/2) # Whitening #

    Academic text (2018)
  • Besides representing the now-familiar millennial convergence of tropics and pole, this image recalls the display of colors associated with the albedo or whitening, the first major transformation in the alchemical process.

    Academic text (1997)
  • In statistics, whitening is crucial for ensuring the data features are uncorrelated and have equal variance.

  • The whitening process helps in improving the performance of machine learning models by removing correlations between features.

Surface Forms

whitening singular
whitenings plural

Morphology

whitening = whiten (semi-transparent) = white + en + ing

The morphological formation from 'whiten' + -ing is regular, but the semantic application is specialized.

Etymology

Whitening in statistics borrows the picture of white ('equal across all parts'), like white noise or white light where no part stands out. So to 'whiten' data is to remove connections between features and make them more even, which helps algorithms work better.