Process mining was born to provide actionable results to companies seeking to understand the reasons behind delays or bottlenecks in their activities. In addition, this tool allows companies to audit and maintain quality control over processes.
Process mining techniques assume a single notion for each case analysis, so they fail to consider different variables that may be present when analyzing an event.
According to the academic Wil van der Aalst, having a single identifier—or a single notion—for each process could lead to biases or loops that do not exist. This is the reason for the emergence of object-centric process mining (OCPM).
What is object-centric process mining?
OCPM aims to avoid convergence and divergence problems by choosing a new record format and by providing new process discovery techniques based on this format.
According to this approach, multiple identifiers that lead to different views on the same process can exist within a case. In addition, an event may be related to various cases (convergence) or the independent and repeated executions of a group of activities may exist within a single case (divergence).
In OCPM, the notion of “case” is abandoned; it is assumed that an event can be related to more than one object.
However, it is important to point out that the classical notion of event registration and the basic process map must be used to represent objects. Similarly, event logs need to contain object descriptions, and process maps can be extended with multi-object representations of activities.
The Inverbis solution
A process can be viewed and analyzed from different perspectives. At the moment, the Inverbis platform provides a standard approach to loading and configuring a record dataset, in which only one field can be selected as an activity. In the future, we will allow users to define more than one field (i.e., multiple mining dimensions) to change the dimension of the analysis when visualizing the process.
Furthermore, we are planning to include a new object-based discovery that focuses on the interactions between the different agents or elements within a process, rather than the actual activity flow. This may be useful for processes that include a considerable number of variables, for which analysis of the relevance of the execution paths of interest is complicated because their frequency of occurrence is very low and homogeneously distributed.
This poses a problem for traditional process mining tools given that the representation of these process flows becomes chaotic despite a potentially clear underlying logic of the interaction between the different agents in the process.
By presenting the point of view of the interaction between objects or agents, the understanding of the key process activities is simplified because agents are represented as nodes of the network and activities are arcs that represent the interaction between the nodes.
In short, this type of approach will facilitate the selection of new analytical and visual representation methods that can be better adapted to the particular characteristics of the processes being studied.
If you want to know more about our solution and how we apply process mining for process improvement, you can click on this link to register and request your demo.