Anomaly Detection for Service-Oriented Business Processes Using Conformance Analysis

A significant challenge for organisations is the timely identification of abnormalities or deviations in their process executions. Abnormalities are generally due to missing vital aspects a possession unwanted behaviour execution. Conformance analysis techniques examine synchronisation between recorded logs and learned models, but exploitation event abnormality detection relatively under-explored area mining. In this paper, we proposed novel technique business execution through extension available conformance techniques. Non-traditional used find correlations discrepancies simulated observed logs. Initially, raw log filtered into two variants, successful failed, based upon outcome instances. Successfully executed instances refer an ideal conduct utilised discover optimal model. Later, model as behavioural benchmark classify failed Abnormal compiled grounded on three dimensions conformance, control flow-based alignment, trace-level alignment event-level alignment. For early predictions, introduced notion lifeline presenting impact varying fitness scores during We applied methodology real-world presented several process-specific improvement measures discussion section.