Skip to main content
probabilistic classifier

probabilistic classifier

6.8
A model that gives the chance of each possible class for an input, for example whether an email is spam
  • noun
  • /ˌprɒbəˈbɪlɪstɪk ˈklæɪsɪfaɪər/
  • Specialized
translation icon : clasificador probabilístico
  • A probabilistic classifier can tell you not only the predicted class but also the likelihood of each possibility.

Examples

  • Training a probabilistic classifier requires models that can output probability distributions, such as logistic regression.

  • Researchers are developing a new probabilistic classifier for medical diagnoses.

  • When dealing with uncertain data, a probabilistic classifier is often preferred over a deterministic one.

  • The probabilistic classifier predicted the outcome with high accuracy.

  • A probabilistic classifier can tell you not only the predicted class but also the likelihood of each possibility.

  • Training a probabilistic classifier requires models that can output probability distributions, such as logistic regression.

  • When dealing with uncertain data, a probabilistic classifier is often preferred over a deterministic one.

Synonyms

classifier
vsprobabilistic classifier
  • Specialized
1 7.1

A computer program that puts data into groups automatically

assigns probabilities across classes rather than picking a single category

Surface Forms

Morphology

probabilistic + classifier

The meaning 'a classifier that assigns probabilities to classes' follows directly from combining 'probabilistic' (based on probability) with 'classifier' (a model that assigns categories). The adjective+noun composition is fully compositional and cross-linguistically common, so a B1 learner who knows both constituents would readily infer the MWE meaning.

Etymology

The term probabilistic classifier brings a simple image: imagine a weather person who does not say just 'rain' or 'no rain' but gives a percent chance for each result. The word probabilistic points to 'chance' and classifier to something that sorts items into groups, so the name means a model that gives the chance for each class instead of 'one fixed answer'.