A surrogate approach to estimate the intrinsic multifractality in financial returns using adaptive network-based fuzzy inference system (ANFIS)

Issue title: Ethical Computational Intelligence for Cyber Market

Guest editors: Oscar Sanjuán Martínez, Giuseppe Fenza and Ruben Gonzalez Crespo

Article type: Research Article

Authors: Wang, Chunxia^{; *}

Affiliations: Department of Basic Science, Wanjiang University of Technology, Ma’anshan, China

Correspondence: [*] Corresponding author. Chunxia Wang, Department of Basic Science, Wanjiang University of Technology, Ma’anshan, 243031, China. Tel.: +8613404150867; Email: enhanlan@126.com.

Abstract: The multifractal behaviors in financial markets results from temporal correlations as well as broad distribution. To evaluate the intrinsic multifractality caused by temporal correlations, surrogate approach is employed under the rank order remapping technique and sign randomization. In contrast to raw multifractality, it is found that intrinsic multifractality is more stable across many years. In this work, we utilize ANFIS model for estimating the intrinsic multifractality in financial returns. Furthermore, the intrinsic multifractality of serial major instruments are highly correlated, which can be served as an index of global market.

Keywords: Multifractality, financial market, surrogate, temporal correlation, ANFIS model

DOI: 10.3233/JIFS-189674

Journal: Journal of Intelligent & Fuzzy Systems, vol. 40, no. 4, pp. 8541-8547, 2021