Process Mining

 View Only

Welcome to the Process Mining Topic Group

  • 1.  Welcome to the Process Mining Topic Group

    IBM TechXchange Speaker
    Posted Thu March 04, 2021 09:23 AM

    What Is Process Mining?

    Process mining sits at the intersection of business process management (BPM) and data mining. While process mining and data mining both work with data, the scope of each dataset differs. Process mining specifically uses event log data to generate process models which can be used to discover, compare, or enhance a given process. The scope of data mining is much broader, and it extends to a variety of data sets. It is used to observe and predict behaviors, having applications within customer churn, fraud detection, and market basket analysis to name a few.

    Process mining takes a more data-driven approach to BPM, which has historically been managed more manually. BPM generally collects data more informally through workshops and interviews, and then uses software to document that workflow as a process map. Since the data that informs these process maps is more qualitative, process mining brings a more quantitative approach to a process problem, detailing the actual process through event data.

    Why Process Mining?

    Increasing sales isn't the only way to generate revenue. Six sigma and lean methodologies also demonstrate how the reduction of operational costs can also increase your return-on-investment (ROI). Process mining helps businesses reduce these costs by quantifying the inefficiencies in their operational models, allowing leaders to make objective decisions about resource allocation. The discovery of these bottlenecks can not only reduce costs and expedite process improvement, but it can also drive more innovation, quality, and better customer retention. However, since process mining is still a relatively new discipline, it still has some hurdles to overcome. Some of those challenges include:

    • Data Quality: Finding, merging and cleaning data is usually required to enable process mining. Data might be distributed over various data sources. It can also be incomplete or contain different labels or levels of granularity. Accounting for these differences will be important to the information that a process model yields.
    • Concept drift: Sometimes processes change as they are being analyzed, resulting in concept drift.

    Want to learn more about IBM Process Mining?  Visit here.

    Please join this Group and return often for more conversation and demos.  And tell us what you think!


    DAVID Jenness