9
IRUS Total
Downloads
  Altmetric

Parameter discovery in unsupervised clustering

File Description SizeFormat 
Preclustering_Short(1).pdfAccepted version259.09 kBAdobe PDFView/Open
Title: Parameter discovery in unsupervised clustering
Authors: Clement, V
Heinis, T
Item Type: Conference Paper
Abstract: Analyzing massive amounts of data and extracting value has become key across different disciplines. A plethora of approaches has been developed to analyze the deluge of data. Using these approaches, however, is not straightforward and many require a priori knowledge of the dataset to set parameters, such as the number of clusters, making their use challenging. Lack of knowledge about the dataset means either that the clustering algorithm has to be run multiple times with different parameters or expensive human intervention and in-depth analysis is required, significantly delaying the analysis and reducing its reproducibility. In this paper, we introduce the idea of simple assumptions about the global distribution of some property of the data leading to local, actionable insights. More specifically, we derive configuration parameters for a clustering method from global distribution properties of a dataset.
Issue Date: 6-Jun-2019
Date of Acceptance: 1-Apr-2019
URI: http://hdl.handle.net/10044/1/75874
DOI: 10.1109/ICDE.2019.00160
ISSN: 1084-4627
Publisher: IEEE
Start Page: 1634
End Page: 1637
Journal / Book Title: 2019 IEEE 35TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE 2019)
Copyright Statement: ©2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Sponsor/Funder: European Research Office
Commission of the European Communities
Funder's Grant Number: 720270
785907
Conference Name: IEEE 35th International Conference on Data Engineering (ICDE)
Keywords: Science & Technology
Technology
Computer Science, Information Systems
Computer Science
Science & Technology
Technology
Computer Science, Information Systems
Computer Science
Publication Status: Published
Start Date: 2019-04-08
Finish Date: 2019-04-11
Conference Place: Macau, PEOPLES R CHINA
Online Publication Date: 2019-06-06
Appears in Collections:Computing