Premium
EVIDENTIAL PROBABILITIES ARE NOT ENOUGH
Author(s) -
Fetzer James H.
Publication year - 1994
Publication title -
computational intelligence
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.353
H-Index - 52
eISSN - 1467-8640
pISSN - 0824-7935
DOI - 10.1111/j.1467-8640.1994.tb00146.x
Subject(s) - citation , psychology , library science , philosophy , information retrieval , computer science
where the double line represents an inconclusive mode of inference (the conclusion can be false even when its premises are true) and the bracketed phrase, “hedged inference,” could, in principle, be replaced by a numerical value representing a quantitative measure of support for C given BK, E . Because Kyburg wants to utilize relative frequencies as the foundation for this conception of evidential probabilities, consider the case in which a sequence of n tosses of coins of a certain irregular structure S has yielded heads H as the outcome with a relative frequency that equals m / n . Knowledge about the outcomes of tosses of this kind presumably would be based upon a series of fallible observations of tosses with coins of that structure. Then a sentence expressing that relative frequency h!F for the outcome heads H in that specific n-membered sequence S would be formalizable as follows: