• The Korean Journal of Applied Statistics
• /
• v.22 no.2
• /
• pp.261-270
• /
• 2009

Current financial crisis triggered by shaky U.S. banking system adds to the emphasis on the importance of the volatility in controlling and understanding financial time series data. The ARCH and GARCH models have been useful in analyzing economic time series volatilities. In particular, multivariate GARCH(MGARCH, for short) provides both volatilities and conditional correlations between several time series and these are in turn applied to computations of hedge-ratio and VaR. In this short article, we try to assess various MGARCH models with respect to the back-testing performances in VaR study. To this end, 14 korean stock prices are analyzed and it is found that MGARCH outperforms rolling window, and BEKK and CCC are relatively conservative in back-testing performance.

• The Journal of Bigdata
• /
• v.4 no.2
• /
• pp.175-183
• /
• 2019

Drop-out issue is one of the challenges of cyber university. There are about 130,000 students enrolled in cyber universities, but the dropout rate is also very high. To lower the dropout rate, cyber universities invest heavily in learning analytics. Some cyber universities analyze the possibility of dropout and actively support students who are more likely to drop out. The purpose of this paper is to identify the learning data affecting the dropout prediction index. As a result of the analysis, it is confirmed that number of lessons(progress), credits, achievement and leave of absence have a significant effect on dropout rate. It is necessary to increase the accuracy of the prediction model through post-test on the student dropout prediction index.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.3-8
• /
• 2021

인공지능 기술이 발달함에 따라 기계번역 기술도 많은 진보를 이루었지만 여전히 기계번역을 통한 번역문 내에는 사람이 교정해야 하는 오류가 다수 포함되어있다. 이렇게 번역 모델에서 생성되는 오류를 교정하는 전문인력의 요구를 경감시키기 위하여 기계번역 사후교정 연구가 등장하였고, 해당 연구는 현재 WMT를 주축으로 활발하게 연구되고 있다. 이러한 사후교정 연구는 최근 도메인 특화 관점에서 주로 연구가 이루어지고 있으며 현재 많은 도메인에서 유의미한 성과를 내고 있다. 하지만 이런 연구들은 기존 번역문의 품질을 얼만큼 향상시켰는가에 초점을 맞출 뿐, 다른 도메인 특화 번역모델의 성능과 비교했을 때 얼마나 뛰어난지는 밝히지 않기 때문에 사후교정 연구가 도메인 특화에서 효과적으로 작용하는지 명확하게 알 수 없다. 이에 본 연구에서는 도메인 특화 번역 모델과 도메인 특화 사후교정 모델간의 성능을 비교함으로써, 도메인 특화에서 사후교정을 통해 얻을 수 있는 실제적인 성능을 검증한다. 이를 통해 사후교정이 도메인 특화 번역모델과 비교했을 때 미미한 수준의 성능을 보임을 실험적으로 확인하였고, 해당 실험 결과를 분석함으로써 향후 도메인특화 사후교정 연구의 방향을 제안하였다.

• Journal of The Korean Association For Science Education
• /
• v.28 no.6
• /
• pp.664-674
• /
• 2008

The purpose of this study was to analyze the effects of the teaching and learning programs for science teachers' scientific hypothesis testing-method invention. Subjects were 30 secondary school science teachers. The potato juice task was administered to the subjects before and after instructional treatments. The four-step strategy including the steps of identifying factors, identifying variables, selecting variables and inventing methods, was applied to the development of the teaching and learning programs for the science teachers. The programs were developed by 6 experts through the R&D method. The results of this study revealed that the number and elaborateness of science teachers' scientific hypothesis testing-methods increased after the instructional treatments. The testing ability of testing-methods invented by science teachers in the posttest was stronger than the one of the methods invented in the pretest. This study also discussed the implications of these findings for teaching and learning in science education.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2000.04a
• /
• pp.890-895
• /
• 2000

본 논문에서는 인공신경망을 이용한 자모 검증을 HMM 기반 온라인 한글인식 시스템에 적용하는 방법론을 제시한다. 본 시스템에서 각각의 자모는 한 개의 HMM 모델과 한 개의 인공신경망 검증기를 갖는다. 자모 검증기는 HMM 네트웍이 생성한 자모 후보 가정을 입력으로 받은 후, 이 가정의 타당성에 대한 사후 확률을 출력한다. 이 사후 확률은 Viterbi 탐색시 탐색 경로에 반영된다. 기존 HMM 시스템의 국소적 특징의 한계를 보완하기 위하여, 한글 자모의 기본획 분석에서 얻어진 구조적, 전역적 특징이 자모 검증기에 사용되었다. 한글 낱자인식에 대한 실험 결과 HMM 기반 인식기에 자모 검증기를 도입함으로서 38.5%의 인식 오류를 줄일 수 있었다.

• Fire Science and Engineering
• /
• v.32 no.3
• /
• pp.123-133
• /
• 2018

This study analyzed the effects influencing the elderly's knowledge on fire safety and practices of fire safety accident prevention after developing a Fire Safety Education Program for the elderly. The program was developed through verification using some experts and a pilot study after selecting the objective, goals, contents, and teaching & learning methods of the program based on the basic survey and literature review. The effects of the program were verified by conducting some non-equivalent control group pre- and post-tests after dividing the targets into the experimental group (the Fire Safety Education Program for the elderly was implemented) and control group (only printed hand-outs were given) among 48 elderly people no younger than 65 years of age. A paired t-test revealed some statistically significant differences between the scores of the pre- and post-tests of the experimental and control groups. ANCOVA found that the effects of the program were significantly higher than those of the printed hand-outs because the post-test scores of the experimental group were higher than those of the control group.

• The Korean Journal of Financial Studies
• /
• v.11 no.1
• /
• pp.157-176
• /
• 2005

본 연구는 KOSPI자산 포트폴리오에 대한 VaR를 다양한 ARCH류 모형을 사용하여 추정하고 이들의 예측능력을 평가하였다. 활용된 모형은 우선 기본적인 GARCH(1,1)모형과 레버리지 효과를 감안한 TGARCH모형, 다양한 ARCH모형을 포괄할 수 있는 PGARCH모형, 변동성의 영속성을 고려한 IGARCH모형이 포함되었다. 모형 상호간의 성과비교에 추가하여 ARCH류 모형에서 수익률예측오차의 분포에 따라서 VaR의 예측성과가 얼마나 차이가 발생하는가를 확인하기 위하여 정규분포와 Student-t분포의 성과를 비교하였다. 마지막으로 VaR 추정시에 조건부평균을 무시하는 관례가 어느정도 타당성이 있는지를 확인하기 위하여 1시차 자기회귀과정에 입각한 조건부 평균을 감안한 결과를 검토하였다. ARCH류 모형에서 모형 설명력은 보다 정교한 모형인 TGARCH모형이나 PGARCH모형이 우월하게 나타났지만, VaR의 예측능력 우월성으로 이어지지는 않았다. Student-t분포를 가정한 경우 VaR모형 사후검증성과는 정규분포를 가정한 경우보다 모든 신뢰수준에서 개선되었으며, 조건부평균의 제거는 Student-t분포 가정하에서는 적합하지 않은 것으로 나타났다. ARCH류 모형에서 가장 단순한 형태인 IGARCH모형의 예측성과가 다른 모형들에 비하여 뒤떨어지지 않으며, 더욱 제약된 형태인 RiskMetrics의 EWMA모형이 사후검증에서 우수한 성과를 보여 단순한 모형의 유용성을 확인시켜주고 있다.

• The Korean Journal of Applied Statistics
• /
• v.28 no.3
• /
• pp.541-559
• /
• 2015

VaR is now widely used as an important tool to evaluate and manage financial risks. In particular, it is important to select an appropriate volatility model for the rate of return of financial assets. In this study, both univariate and multivariate models are considered to evaluate VaR of the portfolio composed of KOSPI, Hang-Seng, Nikkei indexes, and their performances are compared through back testing techniques. Overall, multivariate models are shown to be more appropriate than univariate models to estimate the portfolio VaR, in particular DCC and ADCC models are shown to be more superior than others.

• The Korean Journal of Applied Statistics
• /
• v.26 no.1
• /
• pp.163-185
• /
• 2013

In this paper, we investigate the need to employ long-memory volatility models in terms of Value-at-Risk(VaR) estimation. We estimate the VaR of the KOSPI returns using long-memory volatility models such as FIGARCH and FIEGARCH; in addition, via back-testing we compare the performance of the obtained VaR with short memory processes such as GARCH and EGARCH. Back-testing says that there exists a long-memory property in the volatility process of KOSPI returns and that it is essential to employ long-memory volatility models for the right estimation of VaR.

• The Korean Journal of Applied Statistics
• /
• v.29 no.4
• /
• pp.753-771
• /
• 2016

Value at Risk (VaR) is widely used as an important tool for risk management of financial institutions. In this paper we discuss estimation and back testing for VaR of the portfolio composed of KOSPI, Dow Jones, Shanghai, Nikkei indexes. The copula functions are adopted to construct the multivariate distributions of portfolio components from marginal distributions that combine extreme value theory and GARCH models. Volatility models with t distribution of the error terms using Gaussian, t, Clayton and Frank copula functions are shown to be more appropriate than the other models, in particular the model using the Frank copula is shown to be the best.