Logic and Probabilistic Models of Risk in Banks, Business and Quality

Citation

Andrew, A.M. (2000), "Logic and Probabilistic Models of Risk in Banks, Business and Quality", Kybernetes, Vol. 29 No. 9/10. https://doi.org/10.1108/k.2000.06729iae.003

Logic and Probabilistic Models of Risk in Banks, Business and Quality

E.D. Solojentsev, V.V. Karasev and V.E. SolojentsevNauka PublishersSt Petersburg1999120 pp.ISBN 5-02-024917-3In Russian

Keywords: Publication, Logic, Probability, Computer modelling, Analysis, Risk

The text presents logic and probabilistic techniques for computer modelling and analytical study of risks of losses in banks and ventures. The authors claim that in a field of credit risk evaluation the logic and probabilistic techniques may guarantee twofold more accuracy and sevenfold more robustness of credit classification than conventional methods. Namely, it is demonstrated theoretically, as well as by example, that now classical techniques of neural network based classification, using tabular representation of "object-attribute" mapping, may lose in precision when competing with techniques offered in this book.

A complex problem of quantitative evaluation of a risk of losses is linked to classification and estimation tasks. A theory of logic and probabilistic modelling of failure risks in the groups of incompatible events is introduced. Learning and identification techniques of designing reflexive logical models from statistical data are discussed. An outline of software implementation of the logic and probabilistic models is presented. The book is aimed at analysts in banks and ventures, students in business schools and postgraduate researchers, all those interested in numerical estimation and analysis of failure risks.

Those interested in exploring the book's results in detail should contact Professor E. Solojentsev at sol@sapr.ipme.ru