Extracting data from the event log can generate valuable information that cannot be obtained using other methods.
Process mining is a technique in which business processes are extracted from information system event logs and analyzed. It is a business process management practice used for the purpose of discovering new processes, comparing the existing process with the workflow model, and improving the process. Extracting data from the event log can generate valuable information that cannot be obtained using other methods.
There are three categories of process mining. The first is the discovery model, so called because it involves the discovery of previously unknown or undocumented processes. This type of data mining is carried out when there is no model for the workflow or when the existing documentation is known to be flawed. The event logs are then scanned for information, which is parsed to recreate the process. Documentation for the process is then created, based on the data extracted from the event logs.
The second type of process mining is the compliance model. The name derives from its purpose of verifying that the workflow in progress conforms to the planned process. Event logs are extracted to find differences between the existing process and the model.
Once located, these differences are analyzed to see if they have improved the process. If such changes are beneficial to the process, the model is revised to include these deviations. Decisions made at process control points are reviewed for the information available at each point and the data that affects those decisions. If such changes are disadvantageous, changes can be made to the existing process to allow it to fit more easily into the model.
The third class of process mining is the extension model. This type of data mining seeks to extend an existing model with an enhancement. Event log data is analyzed for possible areas of improvement in the model structure. Bottlenecks, for example, can be checked for possible alternative routes in the workflow.
Process mining presents difficulties. Some tasks are invariably hidden from event logs and cannot be mined. They can be reconstructed by careful analysis of the visible tasks, but not always. Therefore, conclusions based solely on information extracted from event logs may be of questionable quality.
Duplicate tasks in the event log also create problems, as there can be multiple activities in the same category or task name. Therefore, it can be difficult to distinguish between tasks with the same name despite having different functions. Other issues include adequate data on decision making, time built into the model, different perspectives, incorrectly recorded data, and simply not enough information. Process mining must be tempered with experience and common sense to overcome these issues when applying this technique.