The framework for spatiotemporal sequential rule mining: Crime data case study

摘要

Sequence pattern and sequential rule mining techniques are used in many application areas. In our research, we look at the possibility of applying these techniques to spatiotemporal data for the use case of crime data mining. For that, we design the framework that coverts data information into spatiotemporal sequences, in order to detect sequential rules. In the paper, we look at the each step of the framework: data normalization, geocoding, building spatiotemporal enumerated entities, building sequences and applying the sequence pattern mining algorithms. We provide the description of the methods that are involved in the framework workflow with their full description jointly with the pseudo code. For the case study, we use publicly available crime data. Based on that data, we design the use case and show how the framework can be used to find patterns and rules among crimes from the spatial and temporal perspectives.