• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.150-150
• /
• 2018

The Numerical Weather Prediction (NWP) models provide information for weather forecasts. The highly nonlinear and complex interactions in the atmosphere are simplified in meteorological models through approximations and parameterization. Therefore, the simplifications may lead to biases and errors in model results. Although the models have improved over time, the biased outputs of these models are still a matter of concern in meteorological and hydrological studies. Thus, bias removal is an essential step prior to using outputs of atmospheric models. The main idea of statistical bias correction methods is to develop a statistical relationship between modeled and observed variables over the same historical period. The Model Output Statistics (MOS) would be desirable to better match the real time forecast data with observation records. Statistical post-processing methods relate model outputs to the observed values at the sites of interest. In this study three methods are used to remove the possible biases of the real-time outputs of the Weather Research and Forecast (WRF) model in Imjin basin (North and South Korea). The post-processing techniques include the Linear Regression (LR), Linear Scaling (LS) and Power Scaling (PS) methods. The MOS techniques used in this study include three main steps: preprocessing of the historical data in training set, development of the equations, and application of the equations for the validation set. The expected results show the accuracy improvement of the real-time forecast data before and after bias correction. The comparison of the different methods will clarify the best method for the purpose of the forecast skill enhancement in a real-time case study.

• Proceedings of the KIEE Conference
• /
• 2004.11b
• /
• pp.91-93
• /
• 2004

Load forecasting is essential in order to supply electrical energy stably and economically in power systems. ANNs have flexibility to predict a nonlinear feature of load profiles. In this paper, we selected just the necessary input variables used in the paper(2) which is based on the phase-space embedding of a load time-series and reviewing others. So only 5 input variables were selected to forecast for spring, fall and winter season and another input considering temperature sensitivity is added during the summer season. The training cases are also selected from all previous data composed training cases of a 7-day, 14-day and 30-day period. Finally, we selected the training case of a 7-day period because it can be used in STLF without sacrificing the accuracy of the forecast. This allows more compact ANNs, smaller training cases. Consequently, test results show that compact neural networks can be forecasted without sacrificing the accuracy.

• Atmosphere
• /
• v.18 no.3
• /
• pp.237-248
• /
• 2008

A thermal index which considers metabolic heat generation of human body is proposed for operational forecasting. The new thermal index, Perceived Temperature (PT), is forecasted using Weather Research and Forecasting (WRF) mesoscale model and validated. Forecasted PT shows the characteristics of diurnal variation and topographic and latitudinal effect. Statistical skill scores such as correlation, bias, and RMSE are employed for objective verification of PT and input meteorological variables which are used for calculating PT. Verification result indicates that the accuracy of air temperature and wind forecast is higher in the initial forecast time, while relative humidity is improved as the forecast time increases. The forecasted PT during 2007 summer is lower than PT calculated by observation data. The predicted PT has a minimum Root-Mean-Square-Error (RMSE) of $7-8^{\circ}C$ at 9-18 hour forecast. Spatial distribution of PT shows that it is overestimated in western region, while PT in middle-eastern region is underestimated due to strong wind and low temperature forecast. Underestimation of wind speed and overestimation of relative humidity have caused higher PT than observation in southern region. The predicted PT from the mesoscale model gives appropriate information as a thermal index forecast. This study suggests that forecasted PT is applicable to the prediction of health warning based on the relationship between PT and mortality.

• Journal of the Korea Convergence Society
• /
• v.9 no.2
• /
• pp.207-215
• /
• 2018

In this study, targeting KOSPI and KOSDAQ listed companies, the relationship between the abnormal investment of companies and analyst earnings forecasts was empirically analyzed. The analysis period of this study spanned from 2003 to 2015 (with that of dependent variables spanning from 2004 to 2016) based on the variables of interest, and among the companies whose earnings per share forecasts were announced by financial analysts, the final sample of 4,917 companies/year that meets the research condition was selected as the target analysis. The results of the empirical analysis are as follows. First, it turned out that the more total abnormal investment, abnormal R&D and abnormal CAPEX investment, the more accurate were analyst earnings forecasts. Second, the more total abnormal investment, abnormal R&D, abnormal CAPEX investment, the more pessimistic analyst earnings forecasts tended to be. Further analysis has shown that these results came more from over investment groups than under investment groups. The results of this study are expected to make additional contributions to the existing studies in that the abnormal investment is considered as a determinant of analyst earnings forecasts.

• Atmosphere
• /
• v.32 no.1
• /
• pp.27-37
• /
• 2022

The Korean Integrated Model (KIM) forecast system was extended to assimilate Horizontal Line-Of-Sight (HLOS) wind observations from the Atmospheric Laser Doppler Instrument (ALADIN) on board the Atmospheric Dynamic Mission (ADM)-Aeolus satellite. Quality control procedures were developed to assess the HLOS wind data quality, and observation operators added to the KIM three-dimensional variational data assimilation system to support the new observed variables. In a global cycling experiment, assimilation of ALADIN observations led to reductions in average root-mean-square error of 2.1% and 1.3% for the zonal and meridional wind analyses when compared against European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecast System (IFS) analyses. Even though the observable variable is wind, the assimilation of ALADIN observation had an overall positive impact on the analyses of other variables, such as temperature and specific humidity. As a result, the KIM 72-hour wind forecast fields were improved in the Southern Hemisphere poleward of 30 degrees.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.5
• /
• pp.95-101
• /
• 2008

Modeling of stock prices forecast has been considered as one of the most difficult problem to develop accurately since stock prices are highly correlated with various environmental conditions including economics and political situation. In this paper, we propose a agent system approach to predict Korea Composite Stock Price Index (KOSPI) using neural network and statistical methods. To minimize mean of prediction error and variation of prediction error, agent system includes sub-agent modules for feature extraction, variables selection, forecast engine selection, and forecasting results analysis. As a first step to develop agent system for KOSPI forecasting, twelve economic indices are selected from twenty two basic standard economic indices using principal component analysis. From selected twelve economic indices, prediction model input variables are chosen again using best-subsets regression method. Two different types data are tested for KOSPI forecasting and the Prediction results showed 11.92 points of root mean squared error for consecutive thirty days of prediction. Also, it is shown that proposed agent system approach for KOSPI forecast is effective since required types and numbers of prediction variables are time-varying, so adaptable selection of modeling inputs and prediction engine are essential for reliable and accurate forecast model.

• Journal of Service Research and Studies
• /
• v.6 no.4
• /
• pp.75-85
• /
• 2016

Coal can be divided into thermal coal and coking coal. The price of thermal coal is basically affected by demand and supply. However, many other factors with regard to economic condition such as exchange rate, economy growth rate also make an influence on the price. This study is targeted to develop a forecast model for thermal coal price by using System Dynamics Method. System dynamics provides results that better reflect the real world by employing an inter-dependent system of variables. This study found out that 8 factors have important influence on the thermal coal price. Most of the data of the variables were acquired from the Bloomberg Database. The period extends to 2 years and 4 months, from May of 2011 to August of 2013. The causal relations among the variables were acquired by regression analysis

• Korean Journal of Organic Agriculture
• /
• v.25 no.3
• /
• pp.555-567
• /
• 2017

This study suggests the yield forecast model for chilli pepper using artificial neural network. For this, we select the most suitable network models for chilli pepper's yield and compare the predictive power with adaptive expectation model and panel model. The results show that the predictive power of artificial neural network with 5 weather input variables (temperature, precipitation, temperature range, humidity, sunshine amount) is higher than the alternative models. Implications for forecasting of yields are suggested at the end of this study.

• Communications for Statistical Applications and Methods
• /
• v.26 no.5
• /
• pp.497-506
• /
• 2019

Forecasting the U.S. employment level is made using machine learning methods of the artificial neural network: deep neural network, long short term memory (LSTM), gated recurrent unit (GRU). We consider the big data of the federal reserve economic data among which 105 important macroeconomic variables chosen by McCracken and Ng (Journal of Business and Economic Statistics, 34, 574-589, 2016) are considered as predictors. We investigate the influence of the two statistical issues of the dimension reduction and time series differencing on the machine learning forecast. An out-of-sample forecast comparison shows that (LSTM, GRU) with differencing performs better than the autoregressive model and the dimension reduction improves long-term forecasts and some short-term forecasts.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.2
• /
• pp.591-598
• /
• 2018

Short term load forecasts complexly affected by socioeconomic factors and weather variables have non-linear characteristics. Thus far, researchers have improved load forecast technologies through diverse techniques such as artificial neural networks, fuzzy theories, and statistical methods in order to enhance the accuracy of load forecasts. Short term load forecast errors for special days are relatively much higher than that of weekdays. The errors are mainly caused by the irregularity of social activities and insufficient similar past data required for constructing load forecast models. In this study, the load characteristics of Lunar New Year's Day holidays well known for the highest error occurrence holiday period are analyzed to propose a load forecast technique for Lunar New Year's Day holidays. To solve the insufficient input data problem, the similarity of the load patterns of past Lunar New Year's Day holidays having similar patterns was judged by Euclid distance. Lunar New Year's Day holidays periods for 2011-2012 were forecasted by the proposed method which shows that the proposed algorithm yields better results than the comprehensive analysis method or the knowledge-based method.