Data Mining and Knowledge Discovery with Emergent Self-Organizing Feature Maps for Multivariate Time Series (1999)
| Venue: | in Kohonen Maps |
| Citations: | 31 - 7 self |
BibTeX
@INPROCEEDINGS{Ultsch99datamining,
author = {A. Ultsch},
title = {Data Mining and Knowledge Discovery with Emergent Self-Organizing Feature Maps for Multivariate Time Series},
booktitle = {in Kohonen Maps},
year = {1999},
pages = {33--46},
publisher = {Elsevier}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Introduction Data Mining aims to discover so far unknown knowledge in large datasets. The most important step thereby is the transition from subsymbolic to symbolic knowledge. Self-Organizing Feature Maps are very helpful in this task. If appropriately used, they exhibit the ability of emergence. I. e. using the cooperation of many neurons, Emergent Feature Maps are able to build structures on a new, higher level. The UMatrix -method visualizes these structures corresponding to structures of the high-dimensional input-space that otherwise would be invisible. A knowledge conversion algorithm transforms the recognized structures into a symbolic description of the relevant properties of the dataset. In chapter two we shortly introduce our approach to Data Mining and Knowledge Discovery, chapter three clarifies the use of Self-Organizing Feature Maps for Data Mining. Chapter four clarifies the use of Feature Maps in order to obtain emergence. In the chapters five a
