Résumé
Spatial time-stamped sequences have information about time and space where events occur. Mining such sequences can bring important insights. However, not all sequences are frequent over an entire dataset. Some are only common in subsets of time and space. This article explains the first tool for mining these sequences in constrained space and time: the GSTSM R package. It allows users to search for spatio-temporal patterns that are not frequent in the entire database, but are dense in restricted time-space intervals. Thus, making it possible to find non-trivial patterns that would not be found using common data mining tools.