A Review on Machine Learning - Spatiotemporal Data Mining: Issues, Tasks and Applications
P. Madhuri, K. Srinivasa Rao, S. K. Yadav · Zenodo (CERN European Organization for Nuclear Research) · 2020
Spatiotemporal data generally involves the state o f an object, an occurrence, or a spatial location over a period. In many application areas, such as traffic control, climate monitoring, and weather prediction, a significa nt amount of spatiotemporal data can be found. These datasets could be gathered in various formats. At different locations at different points of time. Because of the complex nature of spatiotemporal objects and their relationshi ps in both spatial and temporal dimensions, this raises many difficulties in the representation, c ollection, interpretation, and mining of such datasets. Sometimes, spatio-temporal data sets are very board and hard to interpret and view. In this paper we propose several data mining tasks such a s association rules, classification clusters that are analysed and tested to discover information from spatiotemporal datasets. System functional criteria are addressed for certain kinds of information discovery. Finally, applications are presented for spatiotempor al data mining.