To prevent spam users, you can only post on this forum after registration, which is by invitation. If you want to post on the forum, please send me a mail (h DOT m DOT w DOT verbeek AT tue DOT nl) and I'll send you an invitation in return for an account.

process model analysis

Hi all,

Can you please advise on how to simplify this model. This is from an event log containing raw access log of a website. I am trying to study how users navigate to the homepage. The model is a petri net from inductive miner.

Thanks.

petrinet.jpg 190.5K

Comments

  • hverbeekhverbeek Posts: 751
    Hi,

    Perhaps filtering the log first may help, but I cannot tell how to filter in a sensible way. As an alternative, you could assign a category to every page (perhaps the website contains sub sites of sorts) and look whether there is a pattern for these categories. Other then that, I would not know, but it seems clear you need to reduce the complexity somehow.

    Are you using the inductive miner with noise, or without noise? Allowing for noise may simplify the Petri net a bit, but I doubt it will not be enough to turn this Petri net into a readable one.

    Kind regards,
    Eric.
  • folasadefolasade Posts: 9
    Thank you @hverbeek, I have tried categorizing, its not much better. How do I pick the most frequent traces, say, the top ten or five most frequent traces.

    Thank you.
    Folasade
  • hverbeekhverbeek Posts: 751
    Hi Folasade,

    A way of doing this is to visualize the event log using the "Explore Event Log (Trace Variants / Searchabel / Sortable)" visualizer from the LogEnhancement package. If this package is not installed yet, please install it using the Package Manager.

    This visualizer shows you the variants sorted on their frequencies, and allows you to select a collection of variants and export the selected variants to a new event log (please right-click the list with the variants to get a menu that offers this export). 

    Kind regards,
    Eric.
  • folasadefolasade Posts: 9
    Thank you, you made my day. This helped me select different clusters to study.
    What is the difference between "trace" and "variant", they look like same to me.
  • hverbeekhverbeek Posts: 751
    Hi Folasade,

    Usually, a trace is understood as the sequence of events belonging to a single case. In contrast, a variant corresponds to the sequence of only the activity names (whatever the activity name may be). If the only information contained by an event is its corresponding activity name, then traces and variants can indeed be regarded as similar.

    Kind regards,
    Eric.
Sign In or Register to comment.