de.ls5.jlearn.algorithms.packs
Interface SelectiveExploration

public interface SelectiveExploration

Selective exploration can be provided to the learning algorithm with two consequences: 1) before adding a new word to the observations, the learning algorithm will invoke the explore method. Only if true is returned the word will actually be added. 2) the hypothesis produced by the learner will be passed through the completeAutomaton method on getResult. The SelectiveExploration can then, e.g., wire missing transitions.

Author:

fh

Method Summary

Automaton	completeAutomaton(Automaton sys) post-processes the selectively explored hypothesis produced by the learning algorithm
boolean	explore(Word accessSequence, Symbol transition) tests if the word "accessSequence + transition" should be used as a prefix in the observations.

Method Detail

explore

boolean explore(Word accessSequence,
                Symbol transition)
                throws LearningException

tests if the word "accessSequence + transition" should be used as a prefix in the observations.

Parameters:

accessSequence - to current state

transition - to be taken from current state

Returns:

true if word should be used, false otherwise

Throws:

LearningException

completeAutomaton

Automaton completeAutomaton(Automaton sys)

post-processes the selectively explored hypothesis produced by the learning algorithm

Parameters:

sys, - the hypothesis of the learner

Returns:

post-processed automaton