• Journal of the Korea Society for Simulation
• /
• v.23 no.3
• /
• pp.19-25
• /
• 2014

Demand forecasting is the basis of management activities including marketing strategy. Especially, the demand of a part is remarkably important in supply chain management (SCM). In the fields of various industries, the part demand usually has the intermittent characteristic. The intermittent characteristic implies a phenomenon that there frequently occurs zero demands. In the intermittent demands, non-zero demands have large variance and their appearances also have stochastic nature. Accordingly, in the intermittent demand forecasting, it is inappropriate to apply the traditional time series models and/or cause-effect methods such as linear regression; they cannot describe the behaviors of intermittent demand. Markov bootstrap method was developed to forecast the intermittent demand. It assumes that first-order autocorrelation and independence of lead time demands. To release the assumption of independent lead time demands, this paper proposes a modified bootstrap method. The method produces the pseudo data having the characteristics of historical data approximately. A numerical example for real data will be provided as a case study.

• Journal of Korea Water Resources Association
• /
• v.53 no.7
• /
• pp.481-491
• /
• 2020

The development of agricultural land at Saemangeum has required a significant increase in agricultural water use. It has been well acknowledged that salinity plays a critical role in the farming system. Therefore, a systematic study in salinity is necessary to better manage agricultural water. This study aims to develop a stochastic salinity simulation model that simultaneously simulates salinities obtained from different layers. More specifically, this study proposed a two-stage Autoregressive Exgeneous (ARX) model within a hierarchical Bayesian modeling framework. We derived posterior distributions of model parameters and further used them to obtain the predictive posterior distribution for salinities at three different layers. Here, the BIC values are used and compared to determine the optimal model from a set of candidate models. A detailed discussion of the model is provided.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.6
• /
• pp.1337-1348
• /
• 2017

In this paper, we develop the high-risk drinking predictive model in Korea using the cross-sectional data from Korea Community Health Survey (2014). We perform the logistic regression analysis, the decision tree analysis, and the neural network analysis using the data mining technique. The results of logistic regression analysis showed that men in their forties had a high risk and the risk of office workers and sales workers were high. Especially, current smokers had higher risk of high-risk drinking. Neural network analysis and logistic regression were the most significant in terms of AUROC (area under a receiver operation characteristic curve) among the three models. The high-risk drinking predictive model developed in this study and the selection method of the high-risk intensive drinking group can be the basis for providing more effective health care services such as hazardous drinking prevention education, and improvement of drinking program.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.21 no.1
• /
• pp.46-56
• /
• 2018

This paper aimed to identify main factors of community characters, which have an effect on the land surface temperature(LST) change and estimate the impacting coefficient(ratio) of factors in a significant level of statistics. Chungcheongbuk-do province was selected and then partitioned into city and county areas for the sake of convenience of modeling. LST time series data and the community character data were developed based on Terra Satellite MODIS data and collected from the National Statistical Office, respectively. By the cause and effect relationship between community characters and LST, regression coefficients were estimated using a penal model. In a panel modeling, LST and community characters were used as a dependent variable and explanatory variables, respectively. Panel modeling analysis was carried out using statistical package STATA14 and one-way fixed effect model was selected as the most suitable model to evaluate the regression coefficients in the study area. The impacting ratio of LST change by any explanatory variable derived from the regression coefficients of the panel model fixed. Impacting ratios for industrial areas, elevation ${\times}$ building, energy usage, average window speed, non-urban management area, agricultural, nature and environmental conservation, average precipitation were 3.746, 2.856, 2.742, 0.553, 0.102, 0.071 and 0.003, respectively.

• The Korean Journal of Applied Statistics
• /
• v.30 no.2
• /
• pp.243-257
• /
• 2017

In this paper, we compare several methods to approximate option prices: Edgeworth expansion, A-type and C-type Gram-Charlier expansions, a method using normal inverse gaussian (NIG) distribution, and an asymptotic method using nonlinear regression. We used two different types of approximation. The first (called the RNM method) approximates the risk neutral probability density function of the log return of the underlying asset and computes the option price. The second (called the OPTIM method) finds the approximate option pricing formula and then estimates parameters to compute the option price. For simulation experiments, we generated underlying asset data from the Heston model and NIG model, a well-known stochastic volatility model and a well-known Levy model, respectively. We also applied the above approximating methods to the KOSPI200 call option price as a real data application. We then found that the OPTIM method shows better performance on average than the RNM method. Among the OPTIM, A-type Gram-Charlier expansion and the asymptotic method that uses nonlinear regression showed relatively better performance; in addition, among RNM, the method of using NIG distribution was relatively better than others.

• Communications for Statistical Applications and Methods
• /
• v.15 no.4
• /
• pp.531-542
• /
• 2008

The unit root test is the major tool for determining whether we use differencing or detrending to eliminate the trend from time series data. Dickey-Fuller test (Dickey and Fuller, 1979) has the low power of test when the sample size is small or the true coefficient of AR(1) process is almost unit root and the Bayesian unit root test has complicated testing procedure. We propose a new unit root testing procedure, which mixed Bayesian approach with the traditional testing procedure. Using simulation studies, our approach showed locally higher powers than Dickey-Fuller test when the sample size is small or the time series has almost unit root and simpler procedure than Bayesian unit root test procedure. Proposed testing procedure can be applied to the time series data that are not observed as process with unit root.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.442-442
• /
• 2022

고해상도 시공간적 격자 형태의 레이더 강수는 돌발홍수(flash flood)와 같은 기상재해에 대비하기 위하여 실시간 예측정보로 활용된다. 그러나 대부분의 레이더 강수는 과소 추정되는 경향이 있어 정량적인 보정 과정인 QPE (Quantitative Precipitation Estimation)가 필요하다. 일반적으로 레이더 강수자료 보정은 지점 관측자료를 활용하지만, 본 연구에서는 지상 강수량 기반의 고해상도 격자 강수자료를 생산하여 레이더 강수자료와 직접적으로 비교하고자 한다. 이에 고도와 지형적 특성을 고려한 PRISM(Precipitation-elevation Regressions on Independent Slopes Model) 방법을 사용하여 고해상도 격자기반의 자료를 생성하였다. PRISM 방법은 고도와 지리정보를 독립변수로 갖는 회귀모형 기반의 기후인자 추정 모형이다. 생산된 고해상도 격자 강수자료와 레이더 강수자료를 QPF (Quantitative Precipitation Forecast) 모델의 입력자료로 사용하여 예측결과를 비교하였다. 해당 QPF 모델은 이류(advection)와 확률론적 섭동(stochastic perturbation)을 기반으로 하며, 강수 앙상블 자료를 생산한다. QPF 모델에 대해 투 트랙(two-track) 방법으로 생산된 예측정보를 통해 레이더 강수자료의 격자별 후처리 보정이 가능할 것으로 판단된다.

• The Korean Journal of Applied Statistics
• /
• v.21 no.2
• /
• pp.183-200
• /
• 2008

Research for efficient management of the National Pension has been emphasized as the current society trends toward aging and low birth rate. In this article, we suggest a statistical model for effective classification and prediction of the reserve for the survivor's pension in Korea. Logistic regression model is incorporated; correct classification rate, and distribution of the posterior probability for the reserve of survivor's pension are investigated and compared with the results from the general logistic models. Assessment of predictive model is also done with lift graph, ROC curve and K-S statistic. We suggest strategies for reducing financial risks in managing and planning the pension as an application of the suggested model.

• The Korean Journal of Applied Statistics
• /
• v.24 no.3
• /
• pp.535-546
• /
• 2011

This study develops statistical models for the binary forecast of Asian dust days over South Korea in the winter season. For this study, we used three kinds of data; the rst one is the observed Asian dust days for a period of 31 years (1980 to 2010) as target values, the second one is four meteorological factors(near surface temperature, precipitation, snowfall, ground wind speed) in the source regions of Asian dust based on the NCEP reanalysis data and the third one is the large-scale climate indices. Four kinds of statistical models(multiple regression models, logistic regression models, decision trees, and support vector machines) are applied and compared based on skill scores(hit rate, probability of detection and false alarm rate).

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.5
• /
• pp.1099-1107
• /
• 2017

In this study, we analyzed the determinants of wages of college graduates by using the data of "2014 Graduates Occupational Mobility Survey" conducted by Korea Employment Information Service. In general, wages contain two complex pieces of information about whether an individual is employed and the size of the wage. However, in many previous researches on wage determinants, sample selection bias tends to be generated by performing linear regression analysis using only information on wage size. We used the Heckman sample selection models for analysis to overcome this problem. The main results are summarized as follows. First, the validity of the Heckman's sample selection model is statistically significant. Male is significantly higher in both job probability and wage than female. As age increases and parents' income increases, both the probability of employment and the size of wages are higher. Finally, as the university satisfaction increases and the number of certifications acquired increased, both the probability of employment and the wage tends to increase.