Intervention strategies for improving business process outcomes: A causal rule mining approach

Event logs contain a wealth of information about the execution of business process instances, called cases, and their outcomes. However, extracting actionable information from event logs is still a challenge. This paper proposes the use of causal rule mining on event logs to identify interventions for cases, targeting the improvement of the case outcomes. The mined causal rules identify from the case context both success causes of cases with desired outcomes as well as failure causes of cases with unwanted outcomes. Criteria for evaluating different interventions are developed. Next, different intervention strategies are proposed and analyzed. The intervention strategies are evaluated on several real-world event logs under different settings. The experiments show that by using intervention strategies based on causal rules, the expected outcome of business processes can be significantly improved. • Causal rules are mined to identify causes for case outcomes in event logs. • Different causal rule-based intervention strategies for running cases are proposed. • Requirements and metrics for the intervention strategies are defined. • The intervention strategies are evaluated on three real-world event logs.