• Journal of the Korean Data and Information Science Society
• /
• v.21 no.5
• /
• pp.831-839
• /
• 2010

Least squares support vector machine (LS-SVM) is a kernel trick gaining a lot of popularities in the regression and classification problems. We use LS-SVM to propose a iterative algorithm for a nonlinear generalized autoregressive conditional heteroscedasticity model in the mean (GARCH-M) model to estimate the mean and the conditional volatility of stock market returns. The proposed method combines a weighted LS-SVM for the mean and unweighted LS-SVM for the conditional volatility. In this paper, we show that nonlinear GARCH-M models have a higher performance than the linear GARCH model and the linear GARCH-M model via real data estimations.

• Journal of Korea Port Economic Association
• /
• v.31 no.1
• /
• pp.1-14
• /
• 2015

The standard GARCH model imposing symmetry on the conditional variance, tends to fail in capturing some important features of the data. This paper, hence, introduces the models capturing asymmetric effect. They are the EGARCH model and the GJR model. We provide the systematic comparison of volatility models focusing on the asymmetric effect of news on volatility. Specifically, three diagnostic tests are provided: the sign bias test, the negative size bias test, and the positive size bias test. This paper shows that there is significant evidence of GARCH-type process in the data, as shown by the test for the Ljung-Box Q statistic on the squared residual data. The estimated unconditional density function for squared residual is clearly skewed to the left and markedly leptokurtic when compared with the standard normal distribution. The observation of volatility clustering is also clearly reinforced by the plot of the squared value of residuals of export volume and values. The unconditional variance of both export volumes and export value indicates that large shocks of either sign tend to be followed by large shocks, and small shocks of either sign tend to follow small shocks. The estimated export volume news impact curve for the GARCH also suggests that $h_t$ is overestimated for large negative and positive shocks. The conditional variance equation of the GARCH model for export volumes contains two parameters ${\alpha}$ and ${\beta}$ that are insignificant, indicating that the GARCH model is a poor characterization of the conditional variance of export volumes. The conditional variance equation of the EGARCH model for export value, however, shows a positive sign of parameter ${\delta}$, which is contrary to our expectation, while the GJR model exhibits that parameters ${\alpha}$ and ${\beta}$ are insignificant, and ${\delta}$ is marginally significant. That indicates that the asymmetric volatility models are poor characterization of the conditional variance of export value. It is concluded that the asymmetric EGARCH and GJR model are appropriate in explaining the volatility of export volume, while the symmetric standard GARCH model is good for capturing the volatility.

• The Korean Journal of Applied Statistics
• /
• v.30 no.1
• /
• pp.169-180
• /
• 2017

Multivariate GARCH models are interested in conditional variances (volatilities) as well as conditional correlations between return time series. This paper is concerned with high-frequency multivariate financial time series from which realized volatilities and realized conditional correlations of intra-day returns are calculated. Existing multivariate GARCH models are reviewed comparatively with the realized volatility via canonical correlations and value at risk (VaR). Korean stock prices are analysed for illustration.

• Journal of Information Technology Applications and Management
• /
• v.25 no.4
• /
• pp.157-169
• /
• 2018

The purpose of this study was to examine the volatility of bitcoin, diagnose if bitcoin are a systematic risk asset, and evaluate their effectiveness by estimating market beta representing systematic risk using GARCH (Generalized Auto Regressive Conditional Heteroskedastieity) model. First, the empirical results showed that the market beta of Bitcoin using the OLS model was estimated at 0.7745. Second, using GARCH (1, 2) model, the market beta of Bitcoin was estimated to be significant, and the effects of ARCH and GARCH were found to be significant over time, resulting in conditional volatility. Third, the estimated market beta of the GARCH (1, 2), AR (1)-GARCH (1), and MA (1)-GARCH (1, 2) models were also less than 1 at 0.8819, 0.8835, and 0.8775 respectively, showing that there is no systematic risk. Finally, in terms of efficiency, GARCH model was more efficient because the standard error of a market beta was less than that of the OLS model. Among the GARCH models, the MA (1)-GARCH (1, 2) model considering non-simultaneous transactions was estimated to be the most appropriate model.

• The Korean Journal of Financial Management
• /
• v.23 no.2
• /
• pp.189-208
• /
• 2006

In this paper we are concerned with estimation of tail related risk measures for heteroscedastic financial time series and VaR limits that VaR tells us nothing about the potential size of the loss given. So we use GARCH-EVT model describing the tail of the conditional distribution for heteroscedastic financial series and adopt Expected Shortfall to overcome VaR limits. The main results can be summarized as follows. First, the distribution of stock return series is not normal but fat tail and heteroscedastic. When we calculate VaR under normal distribution we can ignore the heavy tails of the innovations or the stochastic nature of the volatility. Second, GARCH-EVT model is vindicated by the very satisfying overall performance in various backtesting experiments. Third, we founded the expected shortfall as an alternative risk measures.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.3
• /
• pp.419-425
• /
• 2010

Kernel machine learning is gaining a lot of popularities in analyzing large or high dimensional nonlinear data. We use this technique to estimate a GARCH model for predicting the conditional volatility of stock market returns. GARCH models are usually estimated using maximum likelihood (ML) procedures, assuming that the data are normally distributed. In this paper, we show that GARCH models can be estimated using kernel machine learning and that kernel machine has a higher predicting ability than ML methods and support vector machine, when estimating volatility of financial time series data with fat tail.

• The Korean Journal of Applied Statistics
• /
• v.29 no.7
• /
• pp.1373-1383
• /
• 2016

The wind energy industry and wind power generation have increased; consequently, the stable supply of the wind power has become an important issue. It is important to accurately predict the wind power with short-term basis in order to make a reliable planning for the power supply and demand of wind power. In this paper, we first analyzed the speed, power and the directions of the wind. The neural network and the time series models (ARMA, ARMAX, ARMA-GARCH, Holt Winters) for wind power generation forecasting were compared based on mean absolute error (MAE). For one to three hour-ahead forecast, ARMA-GARCH model was outperformed, and the neural network method showed a better performance in the six hour-ahead forecast.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.6
• /
• pp.1041-1049
• /
• 2010

In Korean professional baseball, attendance is the largest source of revenue for development of professional baseball and the highest concern of professional baseball teams. So, if there is demand forecasting model, it will be helpful for pennant chasers to work out the strategies for drawing attendance. For this reason, this research intends to suggest the model which estimates Korean professional baseball's attendance and uses all usable variables which have an effect on attendance in limited circumstances. We supposed that dependent variable is attendance as well as several independent variables and error term are homoscedastic variance. And then, we compared the models which assume conditional heteroscedastic variance like GARCH and EGARCH with GARCH-t models which use the assumption that error term's distribution follows student-t distribution. In result of that, we could confirm that the models which were made by using GARCH(1,1)-t made estimates the most accurately among the several models considered.

• Communications for Statistical Applications and Methods
• /
• v.16 no.5
• /
• pp.791-801
• /
• 2009

Value-at-Risk(VaR) has been used as an important tool to measure the market risk. However, the selection of the VaR models is controversial. This paper proposes VaR forecast combinations using support vector machine quantile regression instead of selecting a single model out of historical simulation and GARCH.

• The Korean Journal of Financial Management
• /
• v.10 no.2
• /
• pp.263-301
• /
• 1993

최근의 연구결과에 의하면 분산이 시간에 따라 변화하여 이분산적(異分散的)이며, 시계열상관(時系列相關)이 존재하는 것으로 나타나고 있다. 일정(一定)한 분산을 가정하여 주식수익률(株式收益率)의 움직임을 설명하는 기존의 모형들은 주식수익률(株式收益率)을 예측하는데 편의(偏倚)(bias)를 가지게 되며, 또한 투자자(投資者)들에게 정확한 위험측정(危險測定)의 수단을 제공하지 못하고 있다. 따라서 본 연구는 우리나라 주식수익률(株式收益率)의 분산이 시간에 따라 변화하는지를 살펴보기 위해 종합주가지수(綜合株價指數) 및 규모별(規模別) 지수(指數)를 사용하여 ARCH 및 GARCH 모형을 추정하였다. 또한 기대수익률(期待收益率)과 조건부(條件附) 분산(分散)사이의 다기간(多期間)(intertemporal) 관계를 ARCH-M 및 GARCH-M 모형을 사용하여 추정하였다. 추정결과는 우리나라 주식시장에도 유의적인 ARCH 및 GARCH 효과, 즉 주식수익률이 매우 이분산적(異分散的)인 것으로 나타났다. 그리고 기대수익률(期待收益率)과 조건부(條件附) 분산(分散)사이의 관계에서 ARCH-M 모형과 GARCH-M 모형의 추정결과가 다르게 나타났으나 전체적으로 유의하지 않는 것으로 나타났다. 이러한 본 연구결과로 조건부(條件附) 분산모형(分散模型)을 통하여 기대수익률(期待收益率) 및 분산(分散)의 움직임을 더욱 잘 파악할 수 있을 것으로 생각되며, 따라서 주식수익률(株式收益率) 및 분산(分散)의 예측에 더 좋은 도구로 활용될 수 있을 것으로 생각된다.