Extraction of Highly Correlated Temporal Event Cluster Recurrence from Spatiotemporal Data

Numerous spatiotemporal datasets are available in databases across several science and technology fields. One common example of spatiotemporal data is a satellite-derived time-series image. In this study, a method for extracting the recurrence of temporal changes highly correlated with a specific time-series subsequence in its spatiotemporal neighborhood is developed using a criterion based on support and confidence for association rules. The method was applied to meteorological satellite images, and the result was visualized as hot spots in the spatiotemporal coordinate system, which enabled the detection of the seasonal migration of the intertropical convergence zone.