Copyright @ IFAC Modeling and Control of Economic
Systems. Klagenfurt. Austria. 2001
ELSEVIER
IFAC
PUBLICATIONS
www.elsevier.com/locatelifac
COMPARA TIVE ANALYSIS OF RESULTS OBTAINED BY UTILIZING
EXPERT SYSTEMS AND NEURAL NETWORKS IN FINANCIAL ANALYSIS
Adrian Kapczynski, Tomasz Zurek
Silesian University of Technology, Faculty of Organization and Management
Abstract: In the context of examining the financial situation of chosen companies using
Du Pont analysis in presented paper a comparative analysis of results obtained while
applying solutions based on Artificial Intelligence is provided. Both main tools, neural
networks and expert systems, are used to examine financial aspects of the given
companies. The results obtained from both tools make it possible to provide advantages
and disadvantages of both tools and finally recommendations of using them under
specified conditions. Copyright © 200} }FAC
Keywords: Expert system, neural network, finance
1. INTRODUCTION
Economic development occurring in the
contemporary world causes many changes in the
organization and management of an enterprise.
Nowadays a key factor of competitive advantage is
information - identification of its source, acquisition.
storage, processing and interpretation. In order to
exist in a tough, highly competitive environment the
company is obliged to perfectly identify and
understand its strengths and weaknesses while having
information concerning the weak points of its
competitors and their behaviour. Based on that the
role of a financial expert in a company is still
growing (Michalski, 1999). Now, the basic goal of an
enterprise may be another financial goal. The
attention of the management is focused primarily on
maximizing the value of a company.
Achieving the mentioned goal guarantees profits for
the owners of capital, the survival of the company in
a highly competitive environment and also provides
advantages connected with satisfying demand
generated by customers.
389
Having existing restrictions (political, economical,
lawful) in mind, maximizing the value of an
enterprise is possible only by practising a policy of
steady development of all aspects of running the
business, leading to an optimization of all key
financial coefficients, not only the return on sales. As
stated above, financial analysis becomes a matter of
utmost importance and remunerativeness a part of it
(Czekaj and Dresler, 1997). One of the most effective
is Du Pont analysis, which shows the influence of a
company's total assets structure, assets turnover on
return referring to sale, total assets and equity
(Sierpinska and Jachna, 1997). Present-day advanced
Information Technology creates tools which can to a
certain degree replace the human expert (Michalski,
2000). Among these tools the wo main groups are
expert systems and neural networks. Both of them
function in a different way; each of them has its own
advantages and drawbacks. Expert systems require
providing detailed knowledge (in the form of rules
and facts), which is then used by the inference engine
to draw conclusions. On the other hand, neural
networks learn through examples, so they usually
require big amounts of data to learn in order to be
properly trained (Tadeusiewicz, 1993). Expert
systems have small capabilities for dealing with