Global investing risk: a case study of knowledge assessment via rough sets |
| |
Authors: | Salvatore Greco Benedetto Matarazzo Roman Slowinski Stelios Zanakis |
| |
Institution: | (1) Department of Computing, Hong Kong Polytechnic University, Hong Kong, China |
| |
Abstract: | This paper presents an application of knowledge discovery via rough sets to a real life case study of global investing risk
in 52 countries using 27 indicator variables. The aim is explanation of the classification of the countries according to financial
risks assessed by Wall Street Journal international experts and knowledge discovery from data via decision rule mining, rather
than prediction; i.e. to capture the explicit or implicit knowledge or policy of international financial experts, rather than
to predict the actual classifications. Suggestions are made about the most significant attributes for each risk class and
country, as well as the minimal set of decision rules needed. Our results compared favorably with those from discriminant
analysis and several variations of preference disaggregation MCDA procedures. The same approach could be adapted to other
problems with missing data in data mining, knowledge extraction, and different multi-criteria decision problems, like sorting,
choice and ranking. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|