UNSUPERVISED LEARNING Of DETEMINISTIC FINIT STATE AUTOMATON FROM EXAMPLES

Deteministic finit state (DFA) automata have emerged as an effective tools for agent modeling applications. The problem of automata learning is to determine a DFA from a series of observation and has recently been studied extensively and a number of algorithms has been proposed. These algorithms can be divided into groups : supervised and unsupervised . In supervised algorithms, we have access to a teacher who give a right answer to our queries, whereas in unsupervised algorithms, the instead of asking questions from teacher, the algorithm rely on latest generated model which changes during its execution. The idea of unsupervised learning an automata through a series of examples have advantageous over supervised version in two important counts: no teacher is needed to response queries and having final model is not necessary. US-L* is an unsupervised method of learning automata which was introduced by D. Carmel at el [1] and is based on L*. To improve the speed of algorithm we modified this method. The proposed method was tested using a series of automaton and randomly generated set of examples and the results are presented.