Rough Sets in Economy and Finance
The Rough Set Theory makes it possible to represent and infer knowledge from incomplete or noisy data, and has attracted much focus of the research community and applications have been found in a wide range of disciplines where knowledge discovery and dat
- PDF / 695,853 Bytes
- 65 Pages / 439.363 x 666.131 pts Page_size
- 33 Downloads / 163 Views
Abstract. The Rough Set Theory makes it possible to represent and infer knowledge from incomplete or noisy data, and has attracted much focus of the research community and applications have been found in a wide range of disciplines where knowledge discovery and data mining are indispensable. This paper provides a detailed review of the currently available literature covering applications of rough sets in the economy and finance. The classical rough set model and its important extensions applied to the economic and financial problems in crucial areas of risk management (business failure, credit scoring), financial market prediction, valuation and portfolio management are described, showing that the rough set theory is an interesting and increasingly popular method employed alongside traditional statistical methods, neural networks and genetic algorithms to support resolution of the most difficult problems in economy and finance. Keywords: soft computing, rough sets, artificial intelligence, risk management, stock market prediction, credit scoring.
1
Introduction
Financial ecosystems are characterized by large amounts of noisy and incomplete information and inherent uncertainty of any forward looking predictions. On the other side, any wrong decision can have potentially catastrophic consequences for the economic well-being of individuals, institutions, nations or even the whole world, as recently shown by the demise of Lehmann Brothers and the Eurozone crisis. It is therefore not surprising that academia, financial industry and its regulatory bodies have ever since been looking for methods able to analyze historical and real time data and infer reliable observations, applicable to a wide variety of problems, ranging from macroeconomic crisis prevention to stock market movement prediction and optimal composition of investment portfolios. Statistics and probability theory traditionally form the basis for formal methods used in economy and finance for data analysis and prediction. The applications range from a self-contained technical analysis of historical stock data J.F. Peters and A. Skowron (Eds.): Transactions on Rough Sets XVII, LNCS 8375, pp. 109–173, 2014. c Springer-Verlag Berlin Heidelberg 2014
110
M. Podsiadlo and H. Rybi´ nski
series1 , which tries to discover buy/sell signals by looking for repeatable trend patterns of price, trading volume and related statistical measures, to sophisticated pricing models, which use complex stochastic processes to model the future uncertainty of price movements. The strong mathematical foundation allows us to deliver predictive models being sophisticated enough to capture many aspects of the modeled reality with statistically measurable precision of the results. However, the generated knowledge has to be considered within the limits of the underlying model of reality and associated assumptions, be it the type of statistical distribution, a pricing model or stochastic process. The multivariate models, typical for the financial environment driven by multiple causal factors, g
Data Loading...
