The origin of Big Data analytics technology “Process Mining” is located in the Netherlands. The Dutch computer science professor Wil van der Aalst already researched systems for workflow management and business process management (BPM) at the TU Eindhoven in the early 2000s . He discovered a gap in the analysis of transactional data: So far, we focused on the analysis of data without assigning them to the evaluation of processes.
Process mining can help companies keep track of their operations. Photo: Elnur – shutterstock.com
Process Mining will close this gap between traditional, model-based process analysis and data-centric analysis techniques such as data mining. Unlike data mining the Process Mining focuses on the use of implicit process knowledge contained in the data – going further than traditional business intelligence solutions that focus primarily on individual metrics, delivering point-in-time results and processes through preselection not holistic.
During this time, ProM developed the first process mining technology as open-source software at the TU Eindhoven which was primarily intended for use in science and research. Since 2011, Process Mining has also been made available to companies – with software solutions that achieve a novel combination of event data and process models.
Equipped for the digital corporate world
In order to meet the challenges of digitization and meet the growing customer demand for a personalized user experience, companies need a very good understanding of their own processes. Process mining can help. Compared to traditional data mining methods, process mining does not start at the data level but at the process level. If necessary, the technology makes every step of the process on a traceable document-based basis. This will reveal deviations and bottlenecks that make a process inefficient – and thus the optimization potential.
Process Mining analyzes logs, which are identified as events, ie activities in the process. These events are in turn assigned to a process instance. The process of a process is mapped as a path to chronological events. In addition (further information) such as information about the role of executing employees, the objects processed in the process and the timestamp of execution can be added. Using these event logs, users can recognize the actual processes in the company and compare them with the target process models and derive measures from them.
Next Generation Process Mining
In times when more than 2.5 billion gigabytes of data are generated every day, it is no longer about collecting, but rather the understanding and use of these gigantic amounts of data. Through the progressive development and the integration of machine learning and artificial intelligence into process mining, it becomes possible to generate recommendations for action and decisions based on huge process data sets.
In machine learning, knowledge is generated from history: the system learns from the process flows, recognizes connections between cause and effect and can recognize and explain deviations. Machine Learning enables “Predictive Process Mining”. Which measures are taken in this specific case is not left entirely to the subjective judgment of the user: Process Mining recognizes patterns in huge datasets, the assessment and implementation of measures in the respective organization is carried out by humans.
In the future, process analysis will increasingly be supported by machine learning (ML) and artificial intelligence. However, the central aspect remains user-friendliness – only in this way is it possible to work in companies without expert knowledge.