ChronoEdgeMiner: A Novel Algorithm for Extracting Frequent Temporal Graphs from Data

The discovery of recurring event patterns is a cornerstone of advanced analytics. While basic insights can be derived from raw event sequences, the true potential is realized when time intervals between events are incorporated. This article proposes a methodology for extracting detailed event graphs (chronicles) from timestamped data, aiming to describe events in a factual and detailed form. Unlike conventional methods, these graphs provide a richer and more precise depiction of event sequences. The aim of this proposal is to extract chronicles in an efficient way. The experimental analysis, considering various databases, demonstrates the proposal’s notable superiority against state-of-the-art algorithms, particularly in runtime and memory requirements.