Detection of Anomalies in the Criminal Proceedings Based on the Analysis of Event Logs

摘要

Process mining makes it possible to solve a task of finding and analyzing deviations in the process. System event logs record information about real process behavior. Weaknesses and errors of a workflow can be found during the analysis of logs. This is especially important in areas associated with significant responsibility and risk. In this paper the focus is on the criminal procedure analysis via process mining methods. A model of this process allows for flexibility only in a strictly regulated framework. However, in practice undesired deviations appear and, therefore, need to be detected and prevented. We adopted conformance checking techniques to determine the anomaly of the trace, taking into account the specifics of the process. We also did clustering of anomaly cases to reveal behavior patterns. They will be helpful for identification of potential causes of such anomalies.