PROBABILITY (BAYESIAN)

Charles François (2004). PROBABILITY (BAYESIAN), International Encyclopedia of Systems and Cybernetics, 2(2): 2629.

Collection	International Encyclopedia of Systems and Cybernetics
Year	2004
Vol. (num.)	2(2)
ID	◀ 2629 ▶
Object type	Epistemology, ontology or semantics, Methodology or model

“The measure of the plausibility of an hypothese or proposition”.

Bayesian probabilities offer a measure of subjective evaluations.

Such probabilities “cannot readily be understood as frequencies, but they have a natural interpretation as indicating the speaker's degree of belief or confidence in the statement, given the available evidence” (W.H. JEFFERYS & J.O. BERGER, 1992, p.66).

Bayesian probabilities are used to quantify a belief or hypothesis related to some situation that cannot be experimentally repeated or produced, in which case no frequencies can be measured.

Of course “… subjective estimates (can) be modified as new information becomes available, leading to the revisions of a priori probabilities” (J.van GIGCH, 1978, p.202).

Generally, new information “more or less”settles within the brackets of the former probabilities evaluation. “More or less”means that the effect of the new information will in most cases lead to narrower brackets of evaluation, i.e. diminishing the spread between upper and lesser values. An interesting practical example has been given by some marine biologists who re-evaluated the rockfish stocks in California (K. SHMIDT, 2003, p. 44-47)

See also