Skip to main content
unsupervised

unsupervised

1 6.4
A method that finds patterns and groups in data without labels
  • adjective
  • /ˌʌnˈsuːpərˌvaɪzd/
  • Specialized
translation icon : no supervisado
  • The supervised and unsupervised machine learning techniques are used for automatic text classification.
  • unsupervised classification
  • unsupervised clustering

So with both unsupervised and supervised learning, in fact, with any sort of neural network, scale matters.

Autoplay Next

Examples

  • The thing is make sure that unsupervised learning work.

  • Now to do this, people often used something called unsupervised learning, where you don't actually teach the machine anything beforehand.

  • And, and you know, this is all unsupervised and directly from the patient's data.

  • In this study, sample outliers were checked using both unsupervised learning and supervised learning methods.

    Academic text (2017)
  • Principal component analysis (PCA), an unsupervised learning method, was performed to check the data quality.

    Academic text (2017)
  • Because global scatterplots only allow pairwise comparisons of individual conditions, we performed an unsupervised cluster analysis to discover distinct expression patterns across all the conditions.

    Academic text (2017)
  • A study on lung adenocarcinoma found that unsupervised DNA microarray analysis predicted death with a sensitivity of 38% and a specificity of 81%.

    Academic text (2003)
  • Clustering methods are unsupervised learning tasks with the purpose of identifying structures in an unlabeled dataset by organizing it into homogeneous groups or clusters.

    Academic text (2017)
  • Unsupervised learning algorithms can automatically identify patterns in unsupervised data without any prior labeling.

  • In machine learning, unsupervised models analyze unlabeled data to find hidden structures and groups.

Synonyms

autodidactic
vsunsupervised
  • Specialized
9.6

A computer system that learns by itself without people telling it the correct answers

focuses on finding patterns in unlabeled data rather than creating training targets

How Labeled

SVM
  • Specialized
1 9.5
unsupervised
  • Specialized
1 6.4

Surface Forms

unsupervised positive

Morphology

unsupervised = supervised (transparent) = un + supervise + ed

The technical ML sense is formed by standard negation of 'supervised' (i.e., 'without supervision/labels'), so the morphological relation is transparent.

Etymology

The word unsupervised is made of un- 'not' and super- 'over' with the idea of 'see', so it keeps the idea of 'not watched'. In machine learning, a helpful way to remember this sense is to think of a computer left alone to find groups or patterns without human labels, so it learns by itself.