Subjective concept of probability

from Wikipedia, the free encyclopedia

The subjective concept of probability understands probability as a measure of the certainty of the personal assessment of a situation. This view is in contrast to the objectivistic concepts of probability such as determinism , propensity or frequentism , in which probability is either the expression of a measurement error or the fundamental physical properties of the world.

Bayesian concept of probability

The most important current of probabilistic subjectivism is Bayesian probability theory . Similar to betting , personal safety can be represented as an odds ratio . Without prior knowledge, the hypothesis and the alternative are equally likely ( principle of indifference : p = 0.5). Can based on previous knowledge, e.g. If, for example, an a priori probability is estimated through previous experiments, this knowledge is included in the calculation of the probability by means of Bayesian statistics . The a priori probabilities can also be estimated on the basis of expert knowledge or intuition .

De Finetti's concept of probability

Bruno de Finetti extended the Bayesian approach. According to this, probability is always an expression of our insufficient information.

Both the Bayesian term and de Finetti's approach make it possible to see probability independently of objective coincidence . This means that statements such as “There used to be life on Mars” can be treated as probability statements, in contrast to the objectivistic approaches (such as the Copenhagen interpretation ), which are based on a physical law that has a tendency ( propensity ) to a certain one Result leads. Frequentistic approaches can only make this assessment if there is the possibility of repeated experiments. However, in the subjectivistic conception of probability, the principle of symmetry is rejected, since in reality an experiment can never be exactly repeated.

See also

Individual evidence

  1. ^ David MacKay: Information Theory, Inference, and Learning Algorithms. ISBN 0-521-64298-1 .
  2. ↑ Bruno de Finetti: Probability Theory . Oldenbourg, Munich 1981.