Long-memory stochastic volatility model calibration using deep neural nets
Masroor, Ahnaf
Content Files
MASROOR-THESIS-2021.pdf
Loading…
Download Files
Loading…
Download Counts (All Files)
Loading…
Edit File
Loading…
Permalink
https://hdl.handle.net/2142/110614
Description
Title
Long-memory stochastic volatility model calibration using deep neural nets
Author(s)
Masroor, Ahnaf
Issue Date
2021-04-29
Director of Research (if dissertation) or Advisor (if thesis)
Chronopoulou, Alexandra
Milenkovic, Olgica
Department of Study
Electrical & Computer Eng
Discipline
Electrical & Computer Engr
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Date of Ingest
2021-09-17T02:34:09Z
Keyword(s)
stochastic volatility models
neural network applications
fast calibration
long-memory models
Heston model
European option pricing
volatility model calibration
Abstract
Widespread use of stochastic volatility models in the financial industry is bottlenecked by the complexity and intractability they present. Since the seminal work in quantitative finance by Black et al. and Merton, the infamous Black-Scholes model has been extensively used in the industry for vanilla and exotic option pricing. Although the model assumes constant volatility which is not observed in the market, the widespread use is sustained due to its closed-form solution for European vanilla option.
However, with the advent of neural networks, stochastic volatility models are becoming increasing tractable. The use of neural networks to learn the expectation function of the underlying stochastic volatility processes for calibration makes application of these more involved stochastic volatility models in the industrial settings practical. This thesis extends this application of neural networks to the calibration of long-memory stochastic volatility (LMSV) models, a class of stochastic volatility models characterized by fractional Brownian motion. The specific challenge with these long-memory models is that they are non-Markovian in nature and simulation can be time-consuming and costly. We show that by using neural networks we can capture these non-Markovian characteristics and quickly calibrate them to ever-evolving market conditions despite their high computational cost.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
MASROOR-THESIS-2021.pdf
