Title :
Iterated logarithmic expansions of the pathwise code lengths for exponential families
Author :
Li, Lei ; Yu, Bin
Author_Institution :
Dept. of Stat., Florida State Univ., Tallahassee, FL, USA
fDate :
11/1/2000 12:00:00 AM
Abstract :
Rissanen´s (1978, 1986, 1989, 1996) minimum description length (MDL) principle is a statistical modeling principle motivated by coding theory. For exponential families we obtain pathwise expansions, to the constant order, of the predictive and mixture codelengths used in MDL. The results are useful for understanding different MDL forms
Keywords :
codes; iterative methods; probability; MDL principle; coding theory; constant order; exponential families; iterated logarithmic expansions; minimum description length principle; mixture codelength; pathwise code lengths; pathwise expansions; predictive codelength; statistical modeling; Costs; Finance; Horses; Information rates; Investments; Minimax techniques; Polynomials; Portfolios; Stochastic processes;
Journal_Title :
Information Theory, IEEE Transactions on
