Het beschrijven van causale modellen met behulp van termen uit de informatietheorie

Samenvatting

S. Nowaks' definition of causality with the concepts of statistical independence & conditional probability finds an analogy in the concept of entropy in the theory of signal transmissions. The use of terms taken from information theory in the description of causal relations seems relevant because of the assymetrical character of the relation causing factor/consequence, & the assymetrical relation "sender/receiver" on which information theory is based. The exercised effect is analogous with the concept of "transmission." A description of causal models by means of terms taken from information theory is analogous (as is shown in some arithmetical examples) to the procedure with r analysis. The distribution of entropy (which is a central term in information theory) is known, so statistical testing becomes possible. H. van de Belt.