An SQL-Based Declarative Process Mining Framework for Analyzing Process Data Stored in Relational Databases

Abstract

Recently, the idea of applying process data analysis over relational databases (DBs) has been investigated in the process mining field resulting into different DB schemas that can be used to effectively store process data coming from Process-Aware Information Systems (PAISs). However, although SQL queries are particularly suitable to check declarative rules over traces stored in a DB, a deep analysis of how the existing instruments for SQL-based process mining can be effectively used for process analysis tasks based on declarative process modeling languages is still missing. In this paper, we present a full-fledged framework based on SQL queries over relational DBs for different declarative process mining use cases, i.e., process discovery, conformance checking, and query checking. The framework is used to benchmark different SQL-based solutions for declarative process mining, using synthetic and real-life event logs, with the aim of exploring their strengths and weaknesses.