Crash forecasting in the Korean stock market based on the log-periodic structure and pattern recognitionopen access
- Authors
- Ko, Bonggyun; Song, Jae Wook; Chang, Woojin
- Issue Date
- Feb-2018
- Publisher
- ELSEVIER SCIENCE BV
- Keywords
- Log-periodicityPrice forecastingDiffusion modelPattern recognitionNon-linear time seriesFinancial market
- Citation
- Physica A: Statistical Mechanics and its Applications, v.492, pp.308 - 323
- Indexed
- SCIE
SCOPUS
- Journal Title
- Physica A: Statistical Mechanics and its Applications
- Volume
- 492
- Start Page
- 308
- End Page
- 323
- URI
- https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/150534
- DOI
- 10.1016/j.physa.2017.09.074
- ISSN
- 0378-4371
- Abstract
- The aim of this research is to propose an alarm index to forecast the crash of the Korean financial market in extension to the idea of Johansen-Ledoit-Sornette model, which uses the log-periodic functions and pattern recognition algorithm. We discover that the crashes of the Korean financial market can be classified into domestic and global crises where each category requires different window length of fitted datasets. Therefore, we add the window length as a new parameter to enhance the performance of alarm index. Distinguishing the domestic and global crises separately, our alarm index demonstrates more robust forecasting than previous model by showing the error diagram and the results of trading performance.
- Files in This Item
-
- Appears in
Collections - 서울 공과대학 > 서울 산업공학과 > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.