Title :
On the bounds of the Titchener T-complexity
Author_Institution :
Dept. of Comput. Sci., Univ. of Auckland, Auckland
Abstract :
Titchenerpsilas T-complexity is a computable complexity measure for finite strings with practical applications in a number of areas ranging from similarity measurement to network event detection. While its lower bound for strings of length L is well known, the corresponding upper bound presents a more difficult problem. Knowledge of this bound is desirable in order to relate computable complexities to notions of information and Shannonpsilas entropy. Beyond practical computability (i.e., short L), a previous result by Titchener, Nicolescu, Staiger, Gulliver, and Speidel based on a special string construction algorithm points at an asymptotic L/ log L bound. This paper revisits this result in the light of a more recently discovered link between the T-complexity computation and the theory of cyclic equivalence classes, confirming some of the assumptions made by Titchener et. al. The paper also offers a non-recursive formulation for their ldquoexhaustion lengthsrdquo and their associated T-complexities.
Keywords :
computational complexity; entropy; equivalence classes; Shannons entropy; Titchener T-complexity; cyclic equivalence classes; finite strings; network event detection; similarity measurement; Application software; Area measurement; Computer networks; Computer science; Data compression; Entropy; Event detection; Frequency estimation; Production; Upper bound;
Conference_Titel :
Communication Systems, Networks and Digital Signal Processing, 2008. CNSDSP 2008. 6th International Symposium on
Conference_Location :
Graz
Print_ISBN :
978-1-4244-1875-6
Electronic_ISBN :
978-1-4244-1876-3
DOI :
10.1109/CSNDSP.2008.4610812