Publication detail

Utilization of Artificial Intelligence for Sensitivity Analysis in the Stock Market

JANKOVÁ, Z. DOSTÁL, P.

Original Title

Utilization of Artificial Intelligence for Sensitivity Analysis in the Stock Market

Type

journal article in Scopus

Language

English

Original Abstract

The main contribution of this paper is to perform sensitivity analysis using artificial intelligence methods on the US stock market using alternative psychological indicators. The Takagi-Sugeno fuzzy model applies investor sentiment represented by VIX index and monitors the impact of economic optimism, political stability and control of the corruption index on the S&P 500 stock index. Alternative psychological indicators have been chosen that have not been explored in the context of stock index performance sensitivity. Investors primarily use fundamental and technical analysis as a source to determine when and what to buy into an investment portfolio. However, psychological factors that may indicate the strength of reaction to the market are often neglected. Fuzzy rules are determined and tested using a neuro-fuzzy inference system and then the rules are reduced by fuzzy clustering to improve performance of ANFIS. The membership function is defined as a Gaussian function because it has the least RMSE value. The sensitivity analysis confirmed that there is a significant impact of the political stability index and the economic optimism index on the S&P 500 performance. Conversely, the sensitivity analysis, unlike the previous study, did not confirm the strong impact of VIX on equity index performance. Results indicate that incorporating psychological indicators in macroeconomic models leads to better supervision and control of the financial markets.

Keywords

artificial intelligence; fuzzy approach; fuzzy logic; sensitivity analysis; sentiment; soft computing; stock market

Authors

JANKOVÁ, Z.; DOSTÁL, P.

Released

31. 10. 2019

Publisher

Mendel University Press

Location

Brno

ISBN

1211-8516

Periodical

Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis

Year of study

67

Number

5

State

Czech Republic

Pages from

1269

Pages to

1283

Pages count

1392

URL

BibTex

@article{BUT159655,
  author="Zuzana {Janková} and Petr {Dostál}",
  title="Utilization of Artificial Intelligence for Sensitivity Analysis in the Stock Market",
  journal="Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis",
  year="2019",
  volume="67",
  number="5",
  pages="1269--1283",
  doi="10.11118/actaun201967051269",
  issn="1211-8516",
  url="https://acta.mendelu.cz/67/5/1269/"
}