Determination of The Dimension of High Dimensional Data Sets: Distance Exponent
by
Aleksandar Stojmirovic

Determination of the dimension of large sets of data plays a very important role in data mining. Many such data sets have their features correlated to some extent, with their `true' dimension being much lower than the number of features would suggest. We define the concept of `Distance Exponent', note its relation to some fractal dimensions and use it to estimate the dimensions of some synthetic sets with simple geometry. It appears that the dimension of our sets is severely underestimated and we investigate the possible causes of this.

Date received: October 23, 2000

Copyright © 2000 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # caek-74.