Fractal Based Data Separation in Data Mining

Authors: Marcel Jiřina, M. Jiřina jr.
Publication year: 2013
Publication type: Konferenční příspěvek (mezinárodní konf.)
Periodical: Proceedings of the The Third International Conference on Digital Information Processing and Communications
Projects: Advanced statistical analysis and non-statistical separation techniques for physical processing detection in data sets sampled by means of elementary particle accelerators.
Bibliographic citation: JIŘINA, Marcel, JIŘINA JR., M. Fractal Based Data Separation in Data Mining. In: Proceedings of the The Third International Conference on Digital Information Processing and Communications. Hong Kong: SDIWC, 2013, S. 287-295. ISBN 978-0-9853483-3-5. Dostupné z: http://sdiwc.net/digital-library/fractal-based-data-separation-in-data-mining
ID: 389175
Online record: View online record

Abstract

The separation of the searched data from the rest is an important task in data mining. Three separation/classification methods are presented. Considering data as points in a metric space, the methods are based on transformed distances of neighbors of a given point in a multidimensional space via a function that uses an estimate of scaling exponent. We shortly describe them and show that transformation function has form of the distance to the scaling exponent power. We also show the efficiency of methods presented on artificial as well as on real-life tasks and compare them with other standard as well as advanced approaches.