• The Korean Journal of Applied Statistics
• /
• v.27 no.7
• /
• pp.1139-1149
• /
• 2014

Multivariate volatility is summarized through canonical correlation analysis (CCA). Along with the standard CCA, non-negative and sparse canonical correlation analysis (NSCCA) is introduced to make sure that volatility coefficients are non-negative and the number of coefficients in the volatility CCA is as small as possible. Various multivariate financial time series are analyzed to illustrate the main contribution of the paper.

• Informatization Policy
• /
• v.18 no.1
• /
• pp.45-54
• /
• 2011

The purpose of this study is to analyze the selection behavior decision model of telecommunications service under the telecommunication environments with joining the WiBro and 3GPP LTE-Advanced service at the sam time. The research results are as follows. First, mobile telecommunication service selection behavior is correlated with each other between WiBro and 3GPP LTE-Advanced service. Second, in the bivariate probit model estimates such as age, income, expenditure and income variable are important decision variables of WiBro and 3GPP LTE-Advanced service selection behavior. Gender and service using years, however, are not an important variables. The single and bivariate probit estimate results show the same estimation results. Finally, the two service selecting predictions by using the bivariate probit estimated results, 28.6% of WiBro and 25.3% of 3GPP LTE-Advanced were expected. Also, chance to join the two telecommunications services simultaneously is 19.3%. Therefore, WiBro selection probability is a little higher than 3GPP LTE-Advanced. The implications of these results, the future telecommunications service selection will be determined by economic factors such as income and expenditure. Thus, to enable the WiBro, policy means for reducing the excessive telecommunications expenditure are needed in the future.

• The Korean Journal of Applied Statistics
• /
• v.32 no.1
• /
• pp.69-81
• /
• 2019

We consider bivariate long range dependent (LRD) time series forecasting using a deep learning method. A long short-term memory (LSTM) network well-suited to time series data is applied to forecast bivariate time series; in addition, we compare the forecasting performance with bivariate fractional autoregressive integrated moving average (FARIMA) models. Out-of-sample forecasting errors are compared with various performance measures for functional MRI (fMRI) data and daily realized volatility data. The results show a subtle difference in the predicted values of the FIVARMA model and VARFIMA model. LSTM is computationally demanding due to hyper-parameter selection, but is more stable and the forecasting performance is competitively good to that of parametric long range dependent time series models.

• Journal of Korea Water Resources Association
• /
• v.51 no.1
• /
• pp.49-60
• /
• 2018

The multivariate regional frequency analysis has many advantages such as an adaption of regional parameters and consideration of a correlated structure of the data. The multivariate regional frequency analysis can provide the broader and more detailed information for the hydrological variables. The multivariate regional frequency analysis has not been attempted to model hydrological variables in South Korea yet. Therefore, it is required to investigate the applicability of the multivariate regional frequency analysis in the modeling of the hydrological variables. The current study investigated the applicability of the homogeneous region delineation and their characteristics in bivariate regional frequency analysis of annual maximum rainfall depth-duration data. The K-medoid method was employed as a clustering method. The discordancy and heterogeneous measures were used to assess the appropriateness of the delineation results. According to the results of the clustering analysis, the employed stations could be grouped into five regions. All stations at three of the five regions led to acceptable values of discordancy measures than the threshold. The stations where have short record length led to the large discordancy measures. All grouped regions were identified as a homogeneous region based on heterogeneous measure estimates. It was observed that there are strong cross-correlations among the stations in the same region.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.39 no.5
• /
• pp.561-568
• /
• 2019

Recently, Korea has suffered from severe droughts due to climate change. Therefore, we need to pay attention to the change of drought risk to develop appropriate drought mitigation measures. In this study, we investigated the changes of hydrologic risk of extreme drought using the current observed data and the projected data according to the RCP 4.5 and 8.5 climate change scenarios. The bivariate frequency analysis was performed for the paired data of drought duration and severity extracted by the threshold level method and by eliminating pooling and minor droughts. Based on the hydrologic risk of extreme drought events Jeonbuk showed the highest risk and increased by 51 % than the past for the RCP 4.5 scenario, while Gangwon showed the highest risk and increased by 47 % than the past for the RCP 8.5 scenario.

• Journal of Korea Water Resources Association
• /
• v.49 no.10
• /
• pp.823-833
• /
• 2016

This study developed a trivariate Copula function based drought frequency analysis model to better evaluate the recent 2014~2015 drought event. The bivariate frequency analysis has been routinely used for the drought variables of interest (e.g. drought duration and severity). However, the recent drought patterns showed that the intensity can be regarded as an important factor which is being characterized by short duration and severe intensity. Thus, we used the trivariate Copula function approach to incorporate the trivariate drought characteristics into the drought frequency analysis. It was found that the return periods based on the trivariate frequency analysis are, in general, higher than the existing bivariate frequency analysis. In addition, this study concludes that the increase in drought frequency claimed by the Gumbel copula function has been overestimated compared to the Student t Copula function. In other words, the selection of copula functions is rather sensitive to the estimation of trivariate drought return periods at a given duration, magnitude and intensity.

• The Korean Journal of Financial Management
• /
• v.17 no.2
• /
• pp.211-227
• /
• 2000

본 연구는 외환위기 이후 1998년 10월부터 2000년 3월까지의 일별 데이터를 사용하여 원/달러 역내시장과 역외시장간의 가격정보 이전에 관한 동조화여부를 실증분석 하였다. 원/달러 역내시장의 가격대용으로 원/달러 현물환율을 사용하였으며, 원/달러 역외시장의 가격대용으로 원/달러 역외선물환율인 NDF 1개월물을 사용하였다. 수익률이 중심이 된 기존의 많은 인과관계 연구들과는 달리 본 연구에서는 환율의 변화율에 대한 그랜져 인과관계 분석과 함께 이변량 GARCH모형을 이용하여 두 시장간에 있어서의 환율의 변화율과 변동성의 인과관계를 분석하였다. 그랜져 인과관계분석 결과 현물환율은 역외선물환율에 대해 강한 선도관계를 가지며 상대적으로 약하지만 역외선물환율 또한 현물환율에 대해 선도관계를 가지는 것으로 나타났다. 본 연구에 사용된 이변량 GARCH모형은 AR(1)-GARCH(1,1)모형으로서 분식 결과를 보면 조건부 변동성이 두 시장간에 상호의존적이며 한 시장의 변화율충격이 다른 시장의 변동성에 영향을 미치는 것이 양 시장간에 유의적으로 나타났다. 이는 현물환시장의 거래정보가 역외선물환시장의 가격형성에 영향을 미치며 역외선물환시장 거래정보 또한 현물환시장으로 이전되어 원/달러 역내시장과 역외시장이 잘 동조화 되어 있다고 말할 수 있다. 즉 정보가 먼저 한 시장에 반영 된 후 다른 시장에 전달되는 정보의 일방 통행적 흐름이 아니라 정보의 반영이 두 시장에서 동시에 이루어지고 정보의 흐름이 양방향으로 이루어짐을 알 수 있다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.201-201
• /
• 2022

본 연구에서는 강우량이 여름에 집중되어있는 우리나라의 강우 특성을 잘 나타낼 수 있는 최적의 확률분포형을 선정하고 해석적 확률모델 (Analytical Probabilistic Model, APM)을 개발하여 유출량을 예측하고자 하였다. 국내 10개 지역인 부산, 춘천, 대구, 대전, 전주, 진주, 서울, 속초, 태백, 원주를 연구 지역으로 설정하였고, 30년 시 단위 강우자료를 지역별 interevent time definition(IETD)을 적용하여 강우 사상으로 그룹화하였다. APM 연구에 일반적으로 사용되는 일변수 지수 분포 이외의 이변수 지수, 감마, 이변수 로그정규 확률밀도함수 (Probability Density Function, PDF)를 강우사상의 특성인 강우량, 강우 지속시간, 무강우 시간의 히스토그램에 적용한 결과, 이 변수 로그정규분포가 우리나라의 강우 특성을 가장 잘 대표하였다. 로그정규분포를 이용하여 APM을 유도하고 유출량을 예측하였다. 예측한 유출량에 대한 빈도분석을 수행하여 Storm Water Management Model (SWMM)의 결과와 비교함으로써 유도한 APM의 적합성을 확인하였다. SWMM의 입력 매개변수 보정을 위해서는 서울 군자 지역에서 관측한 실제 강우량 및 유출량 자료를 사용하였다. 로그정규분포로 유도한 APM과 SWMM의 빈도분석 결과를 비교하였을 때 초과 확률과 재현주기 모두 매우 유사한 결과를 나타내었음을 확인하였다.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.11 no.1
• /
• pp.155-166
• /
• 2008

The quantitative bivariate spatial pattern analysis was applied for the water quality and nutrients data of Masan Bay, and for this analysis Pearson's r as aspatial correlation measurement, Moran's I as spatial association measurement and L index as integration of aspatial and spatial measurement methods were used. To understand the aspatial and spatial characteristics implicated in L index, Pearson's r as well as Moran's I were classified into 3 types respectively, and Pearson's r and Moran's I were combined with 9 types, and also quantile of L index value was used for each of those 9 types. Finally, these types were defined as 5 groups having not overlapped L index range. According to the application result of L index groups, bivariate water quality and nutrients showed no aspatial correlation regardless of spatial association in February and July, but they showed aspatial correlation having clustered spatial pattern in May and November. The result of this study providing the guideline for the interpretation of aspatial correlation and spatial association using L index is expected to be helpful for the marine environment pattern analysis using quantitative index for further study.

• Journal of Korea Water Resources Association
• /
• v.50 no.11
• /
• pp.745-758
• /
• 2017

The copula-based models have been successfully applied to hydrological modeling including drought frequency analysis and time series modeling. However, uncertainty estimation associated with the parameters of these model is not often properly addressed. In these context, the main purposes of this study are to develop the Bayesian inference scheme for bivariate copula functions. The main applications considered are two-fold: First, this study developed and tested an approach to copula model parameter estimation within a Bayesian framework for drought frequency analysis. The proposed modeling scheme was shown to correctly estimate model parameters and detect the underlying dependence structure of the assumed copula functions in the synthetic dataset. The model was then used to estimate the joint return period of the recent 2013~2015 drought events in the Han River watershed. The joint return period of the drought duration and drought severity was above 100 years for many of stations. The results obtained in the validation process showed that the proposed model could effectively reproduce the underlying distribution of observed extreme rainfalls as well as explicitly account for parameter uncertainty in the bivariate drought frequency analysis.