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
|