1 The table below contains some statistics about all the different parameter configurations we ran learnlib with.
2 All except
\emph{RivestSchapire
} using the Random test method result in the correct model being learned.
3 When
\emph{WMethod
} is selected as the testing method
\emph{RivestSchapire
} is also able to learn the correct model.
4 \emph{WMethod
} does however increase the time needed to learn the model significantly, when a different learner is used there is no reason not to use the Random testing method.
6 \begin{longtable
}{| l | l | l | c | c | c |
}
7 \caption{Learning parameters and resulting model size.
} \\
\hline
8 Alphabet & Method & Test method & States & Time \\
\hline \hline
9 small & LStar & Random &
2 &
12 sec \\
\hline
10 small & TTT & Random &
2 &
5 sec \\
\hline
11 small & RivestSchapire & Random &
2 &
6 sec \\
\hline
12 small & KearnsVazirani & Random &
2 &
5 sec \\
\hline
13 small & LStar & WMethod &
2 &
35 sec \\
\hline
14 small & TTT & WMethod &
2 &
32 sec \\
\hline
15 small & RivestSchapire & WMethod &
2 &
33 sec \\
\hline
16 small & KearnsVazirani & WMethod &
2 &
33 sec \\
\hline
18 partial & LStar & Random &
4 &
18 sec \\
\hline
19 partial & TTT & Random &
4 &
16 sec \\
\hline
20 partial & RivestSchapire & Random &
4 &
13 sec \\
\hline
21 partial & KearnsVazirani & Random &
4 &
22 sec \\
\hline
22 partial & LStar & WMethod &
4 &
384 sec \\
\hline
23 partial & TTT & WMethod &
4 &
390 sec \\
\hline
24 partial & RivestSchapire & WMethod &
4 &
384 sec \\
\hline
25 partial & KearnsVazirani & WMethod &
4 &
383 sec \\
\hline
27 full & LStar & Random &
5 &
44 sec \\
\hline
28 full & TTT & Random &
5 &
25 sec \\
\hline
29 full & RivestSchapire & Random &
4 &
12 sec \\
\hline
30 full & KearnsVazirani & Random &
5 &
19 sec \\
\hline
31 full & LStar & WMethod &
5 &
2666 sec \\
\hline
32 full & TTT & WMethod &
5 &
2632 sec \\
\hline
33 full & RivestSchapire & WMethod &
5 &
2638 sec \\
\hline
34 full & KearnsVazirani & WMethod & - & - \\
\hline