Izvorni znanstveni članak

A TOOL FOR SUPPORT OF THE KDD PROCESS

Jan Paralič orcid.org/0000-0002-4603-0411 ; Department of Cybernetics and Artificial Intelligence, University of Technology Košice, Košice, Slovak Republic
Peter Bednar ; Department of Cybernetics and Artificial Intelligence, University of Technology Košice, Košice, Slovak Republic

Sažetak

This paper presents basic ideas and results of the GOAL project focusing on its knowledge discovery part. Within this project a KDD (Knowledge Discovery in Databases) package [7] has been designed and implemented. In this paper motivation, architecture, functionality and one two of the implemented DM (data mining) modules of the KDD Package are described in greater detail. KDD Package supports the whole KDD process [10] starting with possibility to connect to various data sources, following with support for data preprocessing, data mining and knowledge visualization. Based on the GOAL pilot applications one of the DM tasks that are supported now, is the classification DM task (by means of C4.5, CN2, RISE and two-way induction approaches). KDD Package is designed as an open system with easy integration of new DM, data preprocessing or visualization modules.

Ključne riječi

knowledge discovery; data mining; databases; knowledge visualization; GIS

Hrčak ID:

78374

URI

https://hrcak.srce.hr/78374

Datum izdavanja:

13.6.2003.

Posjeta: 1.366 *