Merging event logs for process mining: A rule based merging method and rule suggestion algorithm

• We developed a method to merge different event logs for process mining.
• We developed an algorithm that suggests to the user how the datasets are related.
• We implemented both in a popular process mining framework (ProM).
• Effectiveness and efficiency of the above was evaluated with controlled and real-life data.
• This facilitates process mining on multiple data sources.

In an inter-organizational setting the manual construction of process models is challenging because the different people involved have to put together their partial knowledge about the overall process. Process mining, an automated technique to discover and analyze process models, can facilitate the construction of inter-organizational process models. This paper presents a technique to merge the input data of the different partners of an inter-organizational process in order to serve as input for process mining algorithms. The technique consists of a method for configuring and executing the merge and an algorithm that searches for links between the data of the different partners and that suggests rules to the user on how to merge the data. Tool support is provided in the open source process mining framework ProM. The method and the algorithm are tested using two artificial and three real life datasets that confirm their effectiveness and efficiency.