<technical report>
Learning One-Variable Pattern Languages in Linear Average Time

Creator
Language
Publisher
Date
Source Title
Vol
Publication Type
Access Rights
Related DOI
Related URI
Relation
Abstract A new algorithm for learning one-variable pattern languages is proposed and analyzed with respect to its average-case behavior. We consider the total learning time that takes into account all operatio...ns till an algorithm has converged to a correct hypothesis. For the expectation it is shown that for almost all meaningful distributions defining how the pattern variable is replaced by a string to generate random samples of the target pattern language this algorithm converges within a constant number of rounds with a total learning time that is linear in the pattern length. Thus, the algorithm is average-case optimal in a strong sense.show more

Hide fulltext details.

gz trcs140.ps gz 172 KB 64  
pdf trcs140 pdf 369 KB 165  

Details

Record ID
Peer-Reviewed
Type
Created Date 2009.04.22
Modified Date 2013.12.24

People who viewed this item also viewed