tex opgeschoond'
[tt2015.git] / a4 / question3.tex
1 The table below contains some statistics about all the different parameter
2 configurations we ran learnlib with. All except \emph{RivestSchapire} using
3 the Random test method result in the correct model being learned. When
4 \emph{WMethod} is selected as the testing method \emph{RivestSchapire} is also
5 able to learn the correct model.
6 \emph{WMethod} does however increase the time needed to learn the model
7 significantly, when a different learner is used there is no reason not to use
8 the Random testing method.
9
10 \begin{table}[H]
11 \begin{tabular}{lllccc}
12 \toprule
13 Alphabet & Method & Test method & States & Time \\
14 \midrule
15 small & LStar & Random & 2 & 12 sec \\
16 small & TTT & Random & 2 & 5 sec \\
17 small & RivestSchapire & Random & 2 & 6 sec \\
18 small & KearnsVazirani & Random & 2 & 5 sec \\
19 small & LStar & WMethod & 2 & 35 sec \\
20 small & TTT & WMethod & 2 & 32 sec \\
21 small & RivestSchapire & WMethod & 2 & 33 sec \\
22 small & KearnsVazirani & WMethod & 2 & 33 sec \\
23
24 partial & LStar & Random & 4 & 18 sec \\
25 partial & TTT & Random & 4 & 16 sec \\
26 partial & RivestSchapire & Random & 4 & 13 sec \\
27 partial & KearnsVazirani & Random & 4 & 22 sec \\
28 partial & LStar & WMethod & 4 & 384 sec \\
29 partial & TTT & WMethod & 4 & 390 sec \\
30 partial & RivestSchapire & WMethod & 4 & 384 sec \\
31 partial & KearnsVazirani & WMethod & 4 & 383 sec \\
32
33 full & LStar & Random & 5 & 44 sec \\
34 full & TTT & Random & 5 & 25 sec \\
35 full & RivestSchapire & Random & 4 & 12 sec \\
36 full & KearnsVazirani & Random & 5 & 19 sec \\
37 full & LStar & WMethod & 5 & 2666 sec \\
38 full & TTT & WMethod & 5 & 2632 sec \\
39 full & RivestSchapire & WMethod & 5 & 2638 sec \\
40 full & KearnsVazirani & WMethod & - & - \\
41 \bottomrule
42 \end{tabular}
43 \caption{Learning parameters and resulting model size.}
44 \end{table}