ETHICAL PRINCIPLES IN BUSINESS AND TRADITIONAL QUESTIONNAIRS

conference paper

English

Discrimination among a set of elements is an important aspect of decision-making algorithms. E.g. the discrimination between good and bad investment portfolios is the key problem of a portfolio optimisation strategy. The paper presents a simple algorithm, which can be used to evaluate a discrimination power. The evaluation makes analysis of large, incomplete, inconsistent and vague data sets much more objective. The original version of the algorithms was used to quantify the discrimination power of sociologic questionnaire and was used to optimise such questionnaires. However not a human being but a prohibitively large set of data or a prohibitively sophisticated knowledge base are studied. Therefore the original algorithm was modified to eliminate active human like participation. For example, a repetition of queries is used to evaluate human respondents limited knowledge, subjectivity etc. On the other hand an expert system gives the same answer irrespective of its frequency of submissions. A simple and easy to understand fuzzy algorithm is presented in details.

Discrimination power, knowledge base quality

FRANKOVÁ, E.; DOHNAL, M.; MARKESOVÁ, J.

1. 1. 2003

Brno

80-214-2445-1

Business and Economic Development in Central and Eastern Europe in the Period of Joining to the European Union

27

34

8