This work proposes a model for the valuation of branch offices of banks based on the rough set theory, which could be used as the basis for a decision-making system for dimensioning strategies of a financial entity. It compares the rough set approach with the competitive discriminant analysis methodology using a common set of data from 421 branches. We pay special attention to data reduction and the creation of decision rules that will allow future branches to be classified. These rules could constitute the basis for the evaluation of the viability of dimensioning strategies for a financial entity. In order to evaluate the predictive capabilities of the decision rules, we present the results of cross-validation tests to evaluate the ability of the model to classify new branches. It appears that the rough sets approach provides a favourable tool for the valuation of branch offices.VALUATION BASED ON ROUGH SETS 189 Section 2 reviews rough sets theory. The relationship between rough sets theory and other classification methods is presented in Section 3. The variables of the information system are discussed in Section 4. In Section 5 a discriminant analysis of the information system formed by 421 branches is analysed. In Section 6 a rough sets analysis is obtained through three steps: variable reduction, rule generation and reclassification, and cross-validation tests.
ROUGH SETS THEORY
Introductory RemarksThe theory of rough sets (RS) proposed by Pawlak (1991) offers a model to deal with imprecise or incomplete information. From the start, the theory has generated great interest among researchers working in the field of automatic learning and knowledge discovery from databases (data mining). In the financial field in particular, subjects such as business failure (McKee, 2000;Slowinski et al., 1999;Zopounidis et al., 1999) and stock market analysis (Bazan et al., 1994;Grzymala-Busse, 1997;Golan, 1995) have been studied.The RS theory is used as a method to analyse data in an information system (IS). We applied this theory to the IS formed by all the branches of a Spanish savings bank. In this analysis the main problems are those related to:• The removal of superfluous variables, in order to obtain minimum subsets of variables (reducts) that assure a satisfactory approximation to the predetermined classification of branches by a decision variable.
190C. PÉREZ-LLERA ET AL.