|
Article information
2002 , Volume 7, ¹ 1, p.15-27
Akume D., Weber G.W.
Cluster algorithms: theory and methods
The aim of this work is to study the suitability of some techniques in the clustering of loan banking contracts. We start by presenting two optimization problems that eventually lead to two clustering algorithms - 'the single-link" and the "k-means". We further compare both methods based on certain assessment techniques, space and time complexity, and conclude that the k-means method is more appropriate in view of forecasting customer behaviour.
[full text] Classificator Msc2000:- *62H30 Classification and discrimination; cluster analysis
- 68T10 Pattern recognition, speech recognition
- 91B06 Decision theory
- 91B28 Finance, portfolios, investment
- 91C20 Clustering
Classificator Computer Science:- *G.3 Probability and Statistics
- I.5 Pattern Recognition
- J.4 Social and Behavioral Sciences (Computer Applications)
Keywords: quality of clustering, objective function, clustering finite set, liquidity planning, single-link hierarchical algorithm, minimum distance method
Author(s):
Akume Daniel
Dr.
Office: Computer Science Department, University of Buea, Cameroon
Address: Cameroon, Buea
E-mail: d_akume@yahoo.ca
Weber GerhardW.
Office: Institute of Applied Mathematics, METU
Address: 64289, Turkey, Ankara
E-mail: gweber@metu.edu.tr
Bibliography link:
Akume D., Weber G.W. Cluster algorithms: theory and methods // Computational technologies. 2002. V. 7. ¹ 1. P. 15-27
|
|
|
|
