Application of Spatiotemporal Association Rules on Solar Data to Support Space Weather Forecasting

Abstract

It is well known that solar energetic phenomena influence the Space Weather, in special those directed to the Earth environment. In this context, the analysis of Solar Data is a challenging task, particularly when are composed of Satellite Image Time Series (SITS). It is a multidisciplinary domain that generates a massive amount of data (several Gigabytes per year). It includes image processing, spatiotemporal characteristics, and the processing of semantic data. Aiming to enhance the SITS analysis, we propose an algorithm called "Miner of Thematic Spatiotemporal Associations for Images" (MiTSAI), which is an extractor of Thematic Spatiotemporal Association Rules (TSARs) from Solar SITS. Here, a description is given about the details of the modern algorithm MiTSAI, which is an extractor of Thematic Spatiotemporal Association Rules (TSARs) from solar Satellite Image Time Series (SITS). In addition, its adaptation to the Space Weather and discussion about the specific use in favor of forecasting activities are presented. Finally, some results of its application specifically to solar flare forecasting are also presented. MiTSAI has to extract interesting new patterns compared with the art-state algorithms.