Please use this identifier to cite or link to this item:
Title: Cluster validity indexes to uncertain data for multi-attribute decision-making datasets
Authors: Chang, TC
Jane, CJ
Chang, M 
Keywords: MADM-index
PBMF-based index
Cluster vector index
Rough set
Issue Date: 2018
Publisher: National Dong Hwa University, Computer Center
Source: Journal of internet technology, 2018, v. 19, no. 2, p. 533-538 How to cite?
Journal: Journal of internet technology 
Abstract: This paper proposes a novel function which is designated as the multi-attribute (MA) index function (derived from the conventional PBMF-index function), is used to evaluate the quality of the clustering solution in terms of the number of clusters assigned to each attribute and the accuracy of the corresponding Rough Set (RS) classification. The MA-index function processes a set of parameter values obtained from the Fuzzy C Mean method, Fuzzy Set theory, and RS theory. The MA-index function is embedded within an iterative procedure designated as a multi-attribute decision-making index method, which optimizes both the number of clusters per attribute in the dataset and the accuracy of the corresponding classification. In other words, the clustering/ classification outcome obtained from the multi-attribute decision making index method provides a suitable basis for the formation of reliable decision-making rules. On the whole, the outcomes reveal that the suggested technique not simply generates a much better clustering efficiency as compared to the single-attribute decision-making (SADM) and also PBMF techniques however additionally supplies a much more trustworthy basis for the removal of decision-making policies.
Description: IEEE International Conference on Applied System Innovation (IEEE ICASI), Fuzhou University, Okinawa, Japan, May 26-Jun 1, 2016
ISSN: 1607-9264
EISSN: 2079-4029
DOI: 10.3966/160792642018031902021
Appears in Collections:Journal/Magazine Article

View full-text via PolyU eLinks SFX Query
Show full item record

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.