Root Cause Analysis
The root cause analysis aims at finding process errors and their causes. In addition, the identified root cause structure is analyzed. These analyses make it possible to determine the proportion of errors found that have the same cause, for example. Process problems can be caused by impact factors such as bottlenecks. In Process Mining, the cause analysis can be automated and thus time can be saved.
Automated Root Cause Analysis
If an automated root cause analysis is used, an algorithm identifies relevant influencing factors and forms rules for the problem causes. To create the rules, the algorithm searches the existing data for conspicuous structures and correlations. From this, rules are derived which have the highest possible coverage with the critical data. The high coverage ensures that the rules are relevant for a large proportion of the data.
In order to limit the algorithm search, the user can select a particular vulnerability to be investigated, e.g. processes with very high cycle times. In this example, the algorithm first checks whether the time deviation is within the defined framework. This sorts out the process cycles that do not last particularly long. The system then analyzes the attributes of the cases that come into question (that is, those that have too long a cycle time). This analysis is used to determine which attributes or attribute combinations frequently occur when processing times are too long.
All of the rules mentioned above are derived from these attributes or combinations. The rules are then prioritized on the basis of their coverage (see graphic). The precision of the rule is also important, since very general rules normally cover a large number of cases. Process analysts can derive appropriate improvement measures from the automatically found cause rules.