SEMI INTERACTIVE METHOD FOR DATA MINING

Lydia Boudjeloud-Assala, François Poulet

2006

Abstract

Usual visualization techniques for multidimensional data sets, such as parallel coordinates and scatter-plot matrices, do not scale well to high numbers of dimensions. A common approach to solve this problem is dimensionality selection. We present new semi-interactive method for dimensionality selection to select pertinent dimension subsets without losing information. Our cooperative approach uses automatic algorithms, interactive algorithms and visualization methods: an evolutionary algorithm is used to obtain optimal dimension subsets which represent the original data set without losing information for unsupervised tasks (clustering or outlier detection) using a new validity criterion. A visualization method is used to present the user interactive evolutionary algorithm results and let him actively participate in evolutionary algorithm search with more efficiency resulting in a faster evolutionary algorithm convergence. We have implemented our approach and applied it to real data set to confirm it is effective for supporting the user in the exploration of high dimensional data sets and evaluate the visual data representation.

Paper Citation

in Harvard Style

Boudjeloud-Assala L. and Poulet F. (2006). SEMI INTERACTIVE METHOD FOR DATA MINING . In Proceedings of the Eighth International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 978-972-8865-42-9, pages 3-10. DOI: 10.5220/0002454600030010

in Bibtex Style

@conference{iceis06,
author={Lydia Boudjeloud-Assala and François Poulet},
title={SEMI INTERACTIVE METHOD FOR DATA MINING},
booktitle={Proceedings of the Eighth International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2006},
pages={3-10},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002454600030010},
isbn={978-972-8865-42-9},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Eighth International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - SEMI INTERACTIVE METHOD FOR DATA MINING
SN - 978-972-8865-42-9
AU - Boudjeloud-Assala L.
AU - Poulet F.
PY - 2006
SP - 3
EP - 10
DO - 10.5220/0002454600030010