GeoKDD reasearch group
The development of methods allowing to perform intelligent data reduction is a central issue in environmental science. Nowadays the availability of massive digital geo-referenced databases led GIS scientists to search for new tools able to make sense of such complexity. The field of Knowledge Discovery in Data (KDD) has proposed a number of tools to deal with this issue, but the adaptation of these tools to the domain of Geosciences needs to be further implemented.
The research and teaching activity of our group deal with the exploration of pattern and relationships in complex geospatial environmental data.
In more details, our research topics and scientific activity cover the following multidisciplinary fields:
- Geomatics, geostatistical and GIS modelling/simulations of environmental data;
- Development and adaptation of Machine Learning algorithms for environmental applications and data mining;
- Pattern recognition, modelling and visualization of natural hazards, socio-economic, and demographic data;
- Spatio-temporal data mining, clustering, and hot spots detection;
- Modelling and assessments of natural hazard and risks (wildfires, landslides, permafrost).
One of the main topic of a current scientific interest deals with the application of Machine Learning algorithms to high-dimensional spatio-temporal data, including topo-climatic and meteorological fields, natural hazards and risks, natural resources assessments (wind fields, solar energy), analysis and recognition of complex spatio-temporal patterns. Within the framework of this research new methods are elaborated and applied for environmental and socio-economic spatio-temporal data.
