SQL Mining: Knowledge Discovery from DML Statements

Process centric knowledge discovery led to the founding of a new discipline called process mining, discipline that uses event logs and has as aim process models' discovery. But, not all information systems are able to record these kind of event logs. Moreover, even the information systems provide event logs, most of the existing process mining techniques focus on the control-flow perspective of processes, therefore neglecting the data-flow. The data-flow of a process relies on the data needed (input data elements) in order to execute each activity and the resulted data (the output data elements) after the execution of each activity. This paper focuses on two directions: a) a novel approach for event log extraction (SQL mining using DML statements) and b) a data-flow visualization of process models. The data-flow visualization is expressed using Product Data Model (PDM), so each activity has input and output data elements, but control-flow process mining algorithms may also be applied.