Header

UZH-Logo

Maintenance Infos

Adaptive density estimation


Mazeika, Arturas; Böhlen, Michael H; Taliun, Andrej (2006). Adaptive density estimation. In: 32nd International Conference on Very Large Data Bases, Seoul, Korea, 12 September 2006 - 15 September 2006, 1191-1194.

Abstract

This demonstration illustrates the APDF tree: an adaptive tree that supports the effective and effcient computation of continuous density information. The APDF tree allocates more partition points in non-linear areas of the density function and fewer points in linear areas of the density function. This yields not only a bounded, but a tight control of the error. The demonstration explains the core steps of the computation of the APDF tree (split, kernel additions, tree optimization, kernel additions, unsplit) and demos the implementation for different datasets.

Abstract

This demonstration illustrates the APDF tree: an adaptive tree that supports the effective and effcient computation of continuous density information. The APDF tree allocates more partition points in non-linear areas of the density function and fewer points in linear areas of the density function. This yields not only a bounded, but a tight control of the error. The demonstration explains the core steps of the computation of the APDF tree (split, kernel additions, tree optimization, kernel additions, unsplit) and demos the implementation for different datasets.

Statistics

Downloads

0 downloads since deposited on 29 May 2012
0 downloads since 12 months

Additional indexing

Item Type:Conference or Workshop Item (Paper), refereed, original work
Communities & Collections:03 Faculty of Economics > Department of Informatics
Dewey Decimal Classification:000 Computer science, knowledge & systems
Language:English
Event End Date:15 September 2006
Deposited On:29 May 2012 09:30
Last Modified:07 Dec 2017 11:33
Publisher:VLDB Endowment
Related URLs:http://aitrc.kaist.ac.kr/~vldb06/
Other Identification Number:merlin-id:6156

Download

Content: Published Version
Filetype: PDF - Registered users only
Size: 594kB