This virtual session aims to bring together a community of researchers and practitioners who apply statistical learning / machine learning approaches to spatial data, or who develop algorithms for and the theory behind such applications. Theoretical work and applied case studies, academic research and reports of practical applications, are all welcome. Possible themes include, but are not limited to:
-Applications of machine learning in geographical research and practice
-New sources of big spatial data
-Machine learning algorithms and predictive statistical methods especially designed for spatial data
-Statistical consequences of non-random distribution of data in space and time, and of spatial auto-correlation
With ever-increasing amounts of data and computing power, statistical learning and machine learning have become widespread approaches to tackle difficult problems in many fields of research. Geography is no exception: Applications range from traditional regression and classification approaches used in remote sensing and predicting species distributions to computer vision algorithms automatically mapping habitats based on georeferenced video data collected by robots. At the same time, spatial data can present special challenges for predictive statistics and machine learning; for example, the data may be clustered in space and time around easily accessible research sites, and spatially auto-correlated. This virtual session will explore opportunities and challenges of statistical and machine learning in all sub-fields of geography.
|Introduction||Andy Stock Columbia University||10||12:00 AM|
To access contact information login