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.

Order of events in generated XES Event Log

Lutz
edited April 2014 in Event Logs

Hello,

i have a question regarding the best practise of generating XES event logs and the ordering of events of one trace. All my events have a timestamp and so far i sorted my events using this timestamp before adding them to a trace. I saw there is a ProM plugin that does this sorting. 

I am not sure if i should sort my events per trace before writing them into the event log and the specification of xes seems a bit unclear to me: it states that the events need to have an order but not how this order needs to be realised. In my case there are two options, one is the ordering by using the timestamp and the second one is the ordering by appearance in the xml structure itself. For example the heuristic miner seems to use the later one and ignores the order that is induced by the timestamp (time extension).

Is there any information about this? And how do/would you generate your event logs?

Best regards, Lutz

Best Answer

  • JBuijs
    Accepted Answer

    Hi,

    Your understanding is correct, an ordering is assumed but not verified.

    You might find this also in the Process Mining book by Wil van der Aalst, but then on a 'general' level for event logs and not XES/OpenXES specific.

    Joos Buijs

    Senior Data Scientist and process mining expert at APG (Dutch pension fund executor).
    Previously Assistant Professor in Process Mining at Eindhoven University of Technology

Answers

  • Hi,

    The order as the events appear in the XES/MXML file is used as the order of the events in the trace.

    There is the (hidden) assumption that if a timestamp of execution if available that the events obey this order (i.e. olders event first, then increasing to newer events).

    Hope this helps.

    Joos Buijs

    Senior Data Scientist and process mining expert at APG (Dutch pension fund executor).
    Previously Assistant Professor in Process Mining at Eindhoven University of Technology
  • Hi Joos,

    thanks for the fast reply! As i understtand you this means that the order of events as it is represented in the xml is always assumed to be the correct order, even if the timestamps suggest otherwise (so an assumption of wrong timestamps(?)).

    In this case the correct way would be to sort the evnts of one trace before the serialization to a xes file.

    Is this correct? And do you know of a paper or book that i could quote on this matter?

Sign In or Register to comment.