Fractional Brownian motion inference of multivariate stochastic differential equations

Ahmed H. Ali, Qutaiba N. Nayef AL-Qazaz


Recently, the financial mathematics has been emerged to interpret and predict the underlying mechanism that generates an incident of concern. A system of differential equations can reveal a dynamical development of financial mechanism across time. Multivariate wiener process represents the stochastic term in a system of stochastic differential equations (SDE). The standard wiener process follows a Markov chain, and hence it is a martingale (kind of Markov chain), which is a good integrator. Though, the fractional Wiener process does not follow a Markov chain, hence it is not a good integrator. This problem will produce an Arbitrage (non-equilibrium in the market) in the predicted series. It is undesired property that leads to erroneous conclusion, as it is not possible to build a mathematical model, which represents the financial phenomenon. If there is Arbitrage (unbalance) in the market, this can be solved by Wick-Itô-Skorohod stochastic integral (renormalized integral). This paper considers the estimation of a system of fractional stochastic differential equations (FSDE) using maximum likelihood method, although it is time consuming. However, it provides estimates with desirable characteristic with the most important consistency. Langevin method can be used to find the mathematical form of the functions of stochastic differential equations. This includes drift and diffusion by estimating conditional mean and variance from the data and finding the suitable function achieves the least error, and then estimating the parameters of the model by numerical optimal solution search method. Data used in this paper consist of three banking sector stock prices including Baghdad Bank (BBOB), the Commercial Bank (BCOI), and the National Bank (BNOI).

Full Text:




  • There are currently no refbacks.

Copyright (c) 2020 Ahmed H. Ali, Qutaiba N. Nayef AL-Qazaz

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

ISSN: 2303-4521

Digital Object Identifier DOI: 10.21533/pen

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License