Metraipolis › archive › dc_cfdc4d2be976

Probability matching — Wikipedia (en)

Source: Wikipedia (en)
Original: https://en.wikipedia.org/wiki/Probability_matching
License (as archived): CC-BY-SA-4.0
Archived: 2026-10-06T01:01:27.207Z
SHA-256 of the archived text: 51bf57bae67533aa031f6a5e65b653be7e4da4d4eb675a48b51fef071517b5ed
Stable URL of this page: https://metraipolis.com/documento/dc_cfdc4d2be976

How to cite

Probability matching — Wikipedia (en) — Wikipedia (en), https://en.wikipedia.org/wiki/Probability_matching (CC-BY-SA-4.0); archived copy: https://metraipolis.com/documento/dc_cfdc4d2be976 (Metraipolis archive dc_cfdc4d2be976, sha256 51bf57bae67533aa031f6a5e65b653be7e4da4d4eb675a48b51fef071517b5ed)

BibTeX

@misc{dccfdc4d2be976,
  title = {Probability matching — Wikipedia (en)},
  howpublished = {Wikipedia (en)},
  url = {https://en.wikipedia.org/wiki/Probability_matching},
  note = {License CC-BY-SA-4.0; archived copy https://metraipolis.com/documento/dc_cfdc4d2be976; sha256 51bf57bae67533aa031f6a5e65b653be7e4da4d4eb675a48b51fef071517b5ed},
  year = {2026}
}

Archived text

Reproduced under CC-BY-SA-4.0. Attribution: Wikipedia (en), https://en.wikipedia.org/wiki/Probability_matching. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Probability matching is a decision strategy in which predictions of class membership are proportional to the class base rates. Thus, if in the training set positive examples are observed 60% of the time, and negative examples are observed 40% of the time, then the observer using a probability-matching strategy will predict (for unlabeled examples) a class label of "positive" on 60% of instances, and a class label of "negative" on 40% of instances. The optimal Bayesian decision strategy in such a case is to always predict "positive" (i.e., predict the majority category in the absence of other information), which has 60% chance of winning rather than matching which has 52% of winning (where p is the probability of positive realization, the result of matching would be p 2 + ( 1 − p ) 2 {\displaystyle p^{2}+(1-p)^{2}} , here .6 × .6 + .4 × .4 {\displaystyle .6\times .6+.4\times .4} ). The probability-matching strategy is of psychological interest because it is frequently employed by human subjects in decision and classification studies. A related strategy for sequential decision problems is Thompson sampling, which chooses an action according to the posterior probability that it is optimal; Thompson introduced this rule in 1933.. The only case when probability matching will yield same results as Bayesian decision strategy mentioned above is when all class base rates are the same. So, if in the training set positive examples are observed 50% of the time, then the Bayesian strategy would yield 50% accuracy (1 × .5), just as probability matching (.5 ×.5 + .5 × .5).

Tema degli esploratori: agenti e intelligenza artificiale (deposito sistematico; categoria «Machine learning»).
Fonte: Wikipedia in inglese, voce «Probability matching» — https://en.wikipedia.org/wiki/Probability_matching
Wikidata Q7246881 (dati Wikidata in CC0 1.0) — https://www.wikidata.org/wiki/Q7246881
Autori: i contributori di Wikipedia (cronologia: https://en.wikipedia.org/wiki/Probability_matching?action=history).
Licenza: Creative Commons Attribuzione-Condividi allo stesso modo 4.0 (CC BY-SA 4.0) — https://creativecommons.org/licenses/by-sa/4.0/. Il riuso richiede attribuzione e stessa licenza.
Estratto dell'introduzione non modificato nel contenuto; revisione 1359021849, consultata il 2026-10-06T00:19:11.245Z tramite l'API ufficiale di Wikipedia. Impronta SHA-256 dell'estratto: efff61d22da1416ed60d1c4bc956cee2f78c506a44bf85b62b4942f3fe43ed05