A Novel Biology inspired Model for Evolutionary Subspace Clustering
Fast growing amounts of application data stored in data bases require automatic pattern detection to enable users to gain an understanding of the gathered information. As a basic data mining task in this research area, subspace clustering attracts increasing attention. The task is to generate meaningful similarity-based groupings in highdimensional and/or noisy data sets which do not exhibit global clusters. The general idea is to locally project the instances into subspaces of the attributes. Whereas until now subspace clustering has been successfully applied to gene expression data and hand writing recognition, in our application, the task is to detect subspace clusters in financial transaction data.
|Authors:||Assent I., Krieger R., Steffens A., Seidl T.|
|Published in:||Proc. Annual Symposium on Nature inspired Smart Information Systems (NiSIS 2006), Puerto de la Cruz, Tenerife|
|Type:||Conference papers (peer reviewed)|
|Research topic:||Data Analysis and Knowledge Extraction|