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.

Heuristic Miner empty output/input sets

lauraUnivpm
edited March 2014 in Process Mining

Hi guys,

I don't understand the presence of some empty I/O sets for tasks in Causal net mined with Heuristic Miner plugin. More precisely, I expect to find such empty sets when considering start/end activities; but what does it mean when a task results to be connected to one, two or..no tasks? For example, by exploiting the "repairExample" excercise, by filtering it (I consider a timeframe which involves just Thursday) I obtain ten instances. Mining them by setting all tresholds to 0 and with the "all task connected" option enalbed, I obtain a causal net where the task "Inform User" has as output set {{},{Repair(Complexe):complete},{Repair (Simple): complete,{Archive Repair: complete}}. What the first output means? Moreover by exploring the annotations, in the output patterns it turns out that there is a trace where "Inform Users" seems not to be followed by any other task: but I have checked the log and it is always followed by another task, it isn't an end task in any trace.

Can anyone help me? Thank you!

Best Answer

  • jribeiro
    Accepted Answer

    Hi!

    By the definition, an empty I/O set of a task X occurs when - in a trace - there is no task Y that can be considered as input/output of X. The evident examples are when X is a start or end task. Remark that a task can be a start/end task even though it has some inputs/outputs (e.g., in some traces the task is the start task, in some other traces it's not). 

    The inputs/outputs of a task are defined by the dependency graph; the joins/splits (i.e., I/O sets) of a task are computed by replaying the log on that dependency graph. Therefore, if some inputs or outputs of a task are missing in the dependency graph then the I/O sets may be empty. These cases usually happen when there low frequent behavior or noise in the data. Choosing the right thresholds may solve this issue.

    Cheers,
    Joel 

     

Sign In or Register to comment.