Fast Feature Selection using Fractal Dimension - Ten Years Later

Authors

  • Caetano Traina Jr. ICMC-USP, Sao Carlos
  • Agma Traina ICMC-USP, Sao Carlos
  • Christos Faloutsos Carnegie Mellon University

Abstract

Here we comment about the works that the original paper published in the 2000 Brazilian Symposium on Databases fostered in the Database and Images Group – GBdI, what by their turn motivated other researches abroad. It is shown that the Fractal Theory is indeed helpful to a large spectrum of activities required to manage large amounts of data. Research derived from the original paper includes speeding up similarity queries, designing of cost models and selectivity estimation for similarity queries, sampling on databases, performing attribute selection, identifying clusters of correlated attributes, as well as correlation clustering on large, high dimensional datasets.

Downloads

Download data is not yet available.

Downloads

Published

2010-05-27