GRR: Systém pro získávání znalostí v geografických datech
Title in English | GRR: Knowledge Discovery System for Geographical data |
---|---|
Authors | |
Year of publication | 2003 |
Type | Article in Proceedings |
Conference | Sborník konference GIS... Ostrava 2003 |
MU Faculty or unit | |
Citation | |
web | http://gis.vsb.cz/Publikace/Sborniky/GIS_Ova/GIS_Ova_2003/Sbornik/Referaty/default.htm |
Field | Informatics |
Keywords | data mining; knowledge discovery; geographical information systems; GRASS; PMML |
Description | GRR is a system for mining in geographical information system GRASS (Geographical Resources Analysis Support System). It is composed of graphical user interface, interface for communication with R system and for communication with GIS. Discovered knowledge is expressed in language Predictive Modeling Markup Language (PMML). The system allows visualization of the knowledge. R is a system for data analysis and graphical presentation of results. R provides wide range of statistical methods and machine learning algorithms, e.g. classification, clustering, linear and nonlinear modelling, association rules etc. We extended R system with inductive logic programming methods which allow generating classification rules in first order logic. Moreover, method for mining frequent patterns was added. The paper describes the structure of GRR system and mention its methods for preprocessing and data mining. The example of work with the system is presented. |
Related projects: |