Диссертация (1137084), страница 21

Просмтор этого файла доступен только зарегистрированным пользователям. Но у нас супер быстрая регистрация: достаточно только электронной почты!

Текст из файла (страница 21)

Note that such a separation of START and COMPLETE events of an activity isconsistent with XES Standard [32]. Moreover, the this standard contains other event types, whichallows for fine recording of the activity life-cycleFigure 3.5: Model view of the event log generator GenaOne of the screens demonstrates to a user the Petri net given as plug-in input (see Figure 3.5).It uses visualisation plug-in from ProM 6 Petrinet package to show the model. A user may usethis screen to specify simulation settings more simply. This is favourable for Petri nets of any size.89Two screens ask a user to pick the initial and the final markings. A user selects initial placesfrom which the initial marking consists. The tool uses final markings to end the simulation.

Oncetokens are added to the final places, the execution ends.Some screens are optional. A series of screens help user to specify output preferences for eachplace with undetermined output.Noise settings may be specified with special screens. Users may specify the noise level andwhich kind of noise will be used. There are two following possibilities:– to use only transitions of the given net,– to add additional artificial transitions.The screen shown in Figure 3.6 allows a user to choose any number of transitions from a modelgiven to appear in event log as noise transitions.

Another screen helps to assign the set of artificialnoise transitions.Figure 3.6: Noise settings of the event log generator GenaAs the tool is also able to record timestamps, there are screens to set corresponding parameters.The screen with time settings allows for specifying an execution time for each model’s transition(including artificial noise transitions), and maximum deviation from this time allowed for noisegenerator.903.2.3Tool EvaluationLet us show the example of event log generation using the presented tool. Figure 3.7 shows aPetri net that consists of 21 transitions, 19 places, and 46 arcs. An event log is generated simulatingthis model using our tool.Figure 3.7: Example Petri netThe following generator settings are used:– we generate a single event log without date/time-stamps;– an event log contains 1000 traces;– each trace can contain maximum 100 events;– a noise level is set to 0, i.e.

we do not add noise.Figure 3.8 shows the summary of generated event log as it can be visualized in ProM 6.Figure 3.8: Example event log summary in ProM 691It can be seen, that the log consists of 1000 traces. Totally there are 21078 events in this log of21 event classes.

The number of event classes is always equal to the number of different transitionlabels.The tool generated 732 variants of a process, i.e. it traversed 732 different routes through themodel from the initial to the final markings. Some of these routes have been traversed severaltimes. The model actually allows for infinite number of process variants, because there are loopsin it. Each trace in the event log contains from 15 up to 57 events, and approximately 17 eventclasses are presented in each trace.

Figure 3.9 shows the first two traces from the generated eventlog in XES format.Consider the second example. To illustrate the noise adding feature, we selected three simpleprocess models. Figure 3.10 shows these process models in its first line (models ), ), and )).For each model we generated sets of event logs with different noise levels. Second line ofFigure 3.10 shows models (), ), and )) discovered using -algorithm [5, 42] from the event logthat has been generated using 5% noise level.

Level of 20% was used when we generated eventlogs from which models in the third line of Figure 3.10 were obtained (models ), ℎ), and )).In the first column of Figure 3.10 (models ), and )) transitions A-st, B-st, and Final wereused as noise transitions. Transitions e1, e3, and e6 were used as noise transitions, when wegenerated event logs for the models in the second column of Figure 3.10 (models ), and ℎ)). Thethird column of Figure 3.10 shows models ( ), and )) in which noise transitions are c and e.Moreover, in all of these cases we use three following artificial noise activities: noise1, noise2,and noise3. Transition skipping was enabled.The -algorithm has been selected intentionally for this illustration. This algorithm in itsbasic version can not successfully deal with noise.

Thus, a reader can see, how the noise additioninfluences the simulation. Other process discovery algorithms can deal with noise more or lesseffectively. For example, Inductive [47] or Heuristics [45] miners, if guided, can ignore infrequentbehaviours, and thus re-discover initial models.We do not show the models discovered from the logs with 0% noise level. For such small cases,these models are totally identical to the original one, if sufficient number of traces is used. Notethat process discovery algorithms may show strange or inappropriate results for a tiny event logs— which contain tens of traces — generated from the complex large models.The complexity of a resulting model highly depends on a structure of original model that wasused for generation, and noise settings.

For example, one can try to add more artificial transitionsto a log than there are transition in the model which is simulated. In such a case, an event log willbe generated, from which one can obtain very sophisticated and chaotic model that has nothingcommon with initial model.A reader may see, that models discovered from event logs with 5% noise level (second rowin Figure 3.10) differ from corresponding original models (top row in Figure 3.10) only in severaltransitions.

Whereas models from the bottom row of Figure 3.10 differ from corresponding original9212345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879<t r a c e><s t r i n g<e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t></ t r a c e><t r a c e><s t r i n g<e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t><e v e n t></ t r a c e>...key=" c o n c e p t : n a m e " v a l u e=" T r a c e ␣ 1 " />< s t r i n g key=" c o n c e p t : n a m e " v a l u e="A" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 4 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 6 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t r 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t r 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 5 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="C" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="B" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="F" /> </ e v e n t>key=" c o n c e p t : n a m e " v a l u e=" T r a c e ␣ 2 " />< s t r i n g key=" c o n c e p t : n a m e " v a l u e="A" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 4 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 6 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 5 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="C" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 1 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="B" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 0 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 1 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 3 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="D" /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e=" t 2 2 " /> </ e v e n t>< s t r i n g key=" c o n c e p t : n a m e " v a l u e="F" /> </ e v e n t>Figure 3.9: Fragment of the event log in XES formatmodels more.

Характеристики

Список файлов диссертации