Authors:
Jiwu Zhao
and
Stefan Conrad
Affiliation:
Heinrich-Heine University, Germany
Keyword(s):
Subspace Clustering, Density, High dimensionality, Entropy.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Big Data
;
Biomedical Engineering
;
Business Analytics
;
Data Analytics
;
Data Engineering
;
Data Management and Quality
;
Data Mining
;
Databases and Information Systems Integration
;
Datamining
;
Enterprise Information Systems
;
Health Information Systems
;
Information Systems Analysis and Specification
;
Knowledge Discovery and Information Retrieval
;
Knowledge Management
;
Knowledge-Based Systems
;
Ontologies and the Semantic Web
;
Sensor Networks
;
Signal Processing
;
Society, e-Business and e-Government
;
Soft Computing
;
Symbolic Systems
;
Web Information Systems and Technologies
Abstract:
Subspace clustering is an extension of traditional clustering that enables finding clusters in subspaces within a data set, which means subspace clustering is more suitable for detecting clusters in high-dimensional data sets. However, most subspace clustering methods usually require many complicated parameter settings, which are almost troublesome to determine, and therefore there are many limitations in applying these subspace clustering methods. In our previous work, we developed a subspace clustering method Automatic Subspace Clustering with Distance-Density function (ASCDD), which computes the density distribution directly in high-dimensional data sets by using just one parameter. In order to facilitate choosing the parameter in ASCDD we analyze the relation of neighborhood objects and investigate a new way of determining the range of the parameter in this article. Furthermore, we will introduce here a new method by applying entropy in detecting potential subspaces in ASCDD, whi
ch evidently reduces the complexity of detecting relevant subspaces.
(More)