Prediction of Volatility of a Composite KOSPI by Bayesian EGARCH Model Averaging Procedure

• 발행기관: 한국자료분석학회
• 저자명: Nan Chen Hea-Jung Kim
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.14 No.4, 1749~1758쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2012.08.30

영문 초록

Volatility is known to be an important characteristic that measures the variation of stock price over time. This paper proposes a prediction method for the volatility of a composite KOSPI via a Bayesian approach. For this, Bayesian model averaging procedure, based on a full Bayesian analysis of EGARCH models, is proposed. The procedure, consists of parameter estimation, model comparison, and volatility prediction. Analysis of each step is fully described, which involves implementation of Markov chain Monte Carlo (MCMC) methodology such as the Metropolis-Hastings algorithm and reversible-jump strategy. Details of MCMC implementation in each step using a composite KOSPI data are provided.

목차

1. Introduction

2. EGARCH model for a composite KOSPI

3. Bayesian Computation

4. Conclusion

Reference

키워드

EGARCH model KOSPI Volatility prediction Markov chain Monte Carlo Model averaging