Machine learning versus econometric jump models in predictability and domain adaptability of index options
- Authors
- Jang, H.; Lee, J.
- Issue Date
- Jan-2019
- Publisher
- Elsevier B.V.
- Keywords
- Bayesian neural network; Domain adaptation; Financial time series; Gaussian process regression; L?vy process; Neural network; Support vector regression
- Citation
- Physica A: Statistical Mechanics and its Applications, v.513, pp.74 - 86
- Journal Title
- Physica A: Statistical Mechanics and its Applications
- Volume
- 513
- Start Page
- 74
- End Page
- 86
- URI
- http://scholarworks.bwise.kr/ssu/handle/2018.sw.ssu/39125
- DOI
- 10.1016/j.physa.2018.08.091
- ISSN
- 0378-4371
- Abstract
- Econometric jump models dealing with key stylized facts in financial option markets have an explicit underlying asset process based on stochastic differential equations. Machine learning models with improved prediction accuracy have elicited considerable attention from researchers in the field of financial application. An intensive empirical study is conducted to compare two methods in terms of model estimation, prediction, and domain adaptation using S&P 100 American/European put options. Results indicated that econometric jump models demonstrate better prediction performance than the best-performing machine learning models, and the estimation results of the former are similar to those of the latter. The former also exhibited significantly better domain adaptation performance than the latter regardless of domain adaptation techniques in machine learning
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Business Administration > School of Finance > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.