| Process mining aims to mine business process models behind event logs generated by enterprise information systems.Process mining is of great significance to process conformance checking,business process optimization,and business process reengineering.The completeness of the event logs plays an important role in process mining.However,due to the high concurrency of business processes,it is difficult for event logs to reach the completeness within a period of time.Therefore,how to mine the process model behind the far from complete event log is a very challenging problem.Most existing approaches are based on direct precedences(a relation between events)extracted from event logs for process mining.Although this type of approach can cope with the incompleteness of event logs to a certain extent,it is still difficult to mine high-quality process models when it is impossible to derive all direct precedences from event logs.For this reason,this paper proposes a block-structured process mining approach for incomplete event logs.The main work of this paper is as follows:(1)A block structure process mining approach based on transitive precedences(another relation between events)that fuses collaborative filtering is proposed.Since the transitive precedences obtained from the same event log imply more process behavior information than the direct precedences,compared with existing approaches,this approach can mine higher quality process models from far incomplete event logs.(2)The proposed method is implemented as a plug-in IM-tpcf of the Pro M platform.The IM-tpcf input is an event log in XES format,and the output is a process model in PNML format.(3)The proposed and existing methods are applied to synthetic and real event logs,and a series of comparative experiments verify the effectiveness and efficiency of the proposed method. |
