Geoinfomatiom technology of spatio-temporal data mining

 

 

Valeri Gitis1, Аlexander Derendyaev1, Pavel Metrikov1, Аlexei Shogin2

 

1 Institute for Information Transmission Problems (IITP) RAS, Moscow, gitis@iitp.ru

2 All-Russian Institute of Scientific and Technical Information (VINITI) RAS, Moscow, alex@viniti.ru

 

 

A new geoinformation technology for spatio-temporal data mining and its implementation in a distributed web-based GIS GeoTime II are discussed. GIS GeoTime II is aimed basically to solve two types of problems: (1) browsing through spatial and spatio-temporal information and estimation of relations between its components and (2) discovery of dependencies in geographic information; forecasting, detection and recognition of specific stationary and dynamical properties of the environment under study. The state-of-the-art approach to solve this kind of problems requires a GIS to be able to support the complex analysis of geographic information, to ensure the high interactivity of the analysis, to integrate data distributed both on network servers and on a user’s PC. Additional requirement is due to the diversity in analysis field of application: the system should be designed in compliance with the open architecture principle and should comprise a basic core and a number of plug-ins distributed both on network servers and on a user’s PC. The usage of plug-ins makes it possible to getbe focused on the subject area, while working with local data ensures their confidentiality.

The research of spatio-temporal processes is commonly based on co-processing of indirectly related multi-disciplinary observations represented by catalogues of events, geographical time series, aerospace monitoring raster data, static linear structures, geologic-geophysical raster maps, etc. In GeoTime II the analysis is based on integrated processing of data that are transformed into the uniform 3-D raster fields with two spatial and one temporal coordinates.

In the presentation we consider the application of our technology to concrete examples: the analysis of seismic networks sensitivity, analysis of earthquake precursors, analysis of earthquakes clusterization, surface-water flow analysis etc.

This work was supported in part by the Program of Presidium of RAS “Fundamentals of scientific distributed data-processing environment based on GRID technology”, the section “Electronic Earth”, and by the RFBR projects no. 06-07-89139, 07-07-12019.