Ko, JonghyeonComuzzi, Marco2023-08-242023-08-2420231867-0202http://dx.doi.org/10.1007/s12599-023-00794-yhttps://dl.gi.de/handle/20.500.12116/42171While a business process is most often executed following a normal path, anomalies may sometimes arise and can be captured in event logs. Event log anomalies stem, for instance, from system malfunctioning or unexpected behavior of human resources involved in a process. To identify and possibly fix these, anomaly detection has emerged recently as a key discipline in process mining. In the paper, the authors present a systematic review of the literature on business process event log anomaly detection. The review aims at selecting systematically studies in the literature that have tackled the issue of event log anomaly detection, classifying existing approaches based on criteria emerging from previous literature reviews, and identifying those research directions in this field that have not been explored extensively. Based on the results of the review, the authors argue that future research should look more specifically into anomaly detection on event streams, extending the number of event log attributes considered to determine anomalies, and producing more standard labeled datasets to benchmark the techniques proposed.Anomaly detection||Business process||Event log quality||Process dataText/Journal Article10.1007/s12599-023-00794-y