id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_ftxcr4leezfmtmjudk2sxht2bq	Elliott Sober	Instrumentalism, Parsimony, and the Akaike Framework	2002	12	.pdf	application/pdf	5195	370	63	Akaike's framework for thinking about model selection in terms of the goal of predictive Scientists often test models whose truth values they already know, and they often often taken to have punctured the instrumentalist balloon with his suggestion that the difference between instrumentalism and realism is nonsubstantive; if true theories are the ones that maximize predictive accuracy, 2. Forster and Sober (1994) describe Akaike's estimated predictive accuracy as a quantity per datum, and so divided the right side of this equation by N, the number of data. is the goal of inference; in addition, he provided a methodology for estimating a model's predictive accuracy. and simplicity are in conflict; Akaike's theorem shows how each contributes to estimating a model's predictive accuracy. is a separate consideration in model selection from fit to data, the justification provided by Akaike's theorem for using simplicity depends on empirical assumptions.	./cache/work_ftxcr4leezfmtmjudk2sxht2bq.pdf	./txt/work_ftxcr4leezfmtmjudk2sxht2bq.txt