Knowledge Discovery From Global Remote Sensing and Climate Data: Results from Supervised and Unsupervised Data Mining

This paper describes results and lessons learned from research activities designed to develop data mining and machine learning methods for remote sensing and Earth science data sets. These data sets are acquired by Earth observing instruments onboard polar orbiting satellites, in-situ observations, and model reanalysis and provide a rich source of information related to the properties and dynamics of the Earth’s land, oceans, and atmosphere. They are characterized by very large volumes, high dimensionality, and possess both spatial and temporal attributes. The result is a suite of extremely complex, high-dimensional, and heterogeneous data sets that present significant analysis challenges.