• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.28 no.6
• /
• pp.319-331
• /
• 2016

This study explores the signal characteristics for different averaging intervals and defines representative verticies for each observatory by criterion of percent rate and variance. The shorter averaging interval shows the higher frequency variation, though the lower percent rate. In the tidal currents, we could hardly find the differences between 60-minute and 20-minute averaging. The newly defined criterion improves reliability of HF-radar data compared with the present reference which deselects the half by percent rate.

• Journal of Environmental Impact Assessment
• /
• v.14 no.2
• /
• pp.55-62
• /
• 2005

This study applied the Neural Network and Fuzzy theory to show water-purity control and preventive measure in water quality forecasting of the future river. This study picked out NAJU and HAMPYUNG as the subject of investigation and used monthly the water quality and the outflow data of KWANGJU2, NAJU, YOUNGSANNPO and HAMPYUNG from 1995 to 1999 to forecast BOD, COD, T-N, T-P water density. The datum from 1995 to 1999 are used for study and that of 2000 are used for verification. To develop model of water quality forecasting, firstly, this research formed Neural Network model and divided Neural Network model into two case - the case of considering lag and not considering. And this study selected optimal Neural Network model through changing the number of hidden layer based on input layer(n) from n to 3n. Through forecasting result, the case without considering lag showed more precise simulated result. Accordingly, this study intended to compare, analyse that Fuzzy model using the method without considering lag with Neural Network model. As a result, this study found that the model without considering lag in Neural Network Network shows the most excellent outcome. Thus this study examined a forecasting accuracy, analyzed result and verified propriety through appling the method of water quality forecasting using Neural Network and Fuzzy Algorithms to the actual case.

• Korean Journal of Ecology and Environment
• /
• v.46 no.1
• /
• pp.1-9
• /
• 2013

Algal blooms in potable water supplies are becoming an increasingly prevalent and serious water quality problem around the world. In addition to precipitating taste and odor problems, blooms damage the environment, and some classes like cyanobacteria (blue-green algae) release toxins that can threaten human health, even causing death. There is a recognized need in the water industry for models that can accurately forecast in real-time algal bloom events for planning and mitigation purposes. In this study, using data for an interconnected system of rivers and reservoirs operated by a New Jersey water utility, various ANN models, including both discrete prediction and classification models, were developed and tested for forecasting counts of three different algal classes for one-week and two-weeks ahead periods. Predictor model inputs included physical, meteorological, chemical, and biological variables, and two different temporal schemes for processing inputs relative to the prediction event were used. Despite relatively limited historical data, the discrete prediction ANN models generally performed well during validation, achieving relatively high correlation coefficients, and often predicting the formation and dissipation of high algae count periods. The ANN classification models also performed well, with average classification percentages averaging 94 percent accuracy. Despite relatively limited data events, this study demonstrates that with adequate data collection, both in terms of the number of historical events and availability of important predictor variables, ANNs can provide accurate real-time forecasts of algal population counts, as well as foster increased understanding of important cause and effect relationships, which can be used to both improve monitoring programs and forecasting efforts.

• Journal of Korean Society for Atmospheric Environment
• /
• v.26 no.6
• /
• pp.666-682
• /
• 2010

The forecasting system for Today's and Tomorrow's PM10 was developed based on the statistical model and the forecasting was performed at 9 AM to predict Today's 24 hour average PM10 concentration and at 5 PM to predict Tomorrow's 24 hour average PM10. The Today's forecasting model was operated based on measured air quality and meteorological data while Tomorrow's model was run by monitored data as well as the meteorological data calculated from the weather forecasting model such as MM5 (Mesoscale Meteorological Model version 5). The observed air quality data at ambient air quality monitoring stations as well as measured and forecasted meteorological data were reviewed to find the relationship with target PM10 concentrations by the regression analysis. The PM concentration, wind speed, precipitation rate, mixing height and dew-point deficit temperature were major variables to determine the level of PM10 and the wind direction at 500 hpa height was also a good indicator to identify the influence of long-range transport from other countries. The neural network, regression model, and decision tree method were used as the forecasting models to predict the class of a comprehensive air quality index and the final forecasting index was determined by the most frequent index among the three model's predicted indexes. The accuracy, false alarm rate, and probability of detection in Tomorrow's model were 72.4%, 0.0%, and 42.9% while those in Today's model were 80.8%, 12.5%, and 77.8%, respectively. The statistical model had the limitation to predict the rapid changing PM10 concentration by long-range transport from the outside of Korea and in this case the chemical transport model would be an alternative method.

• Atmosphere
• /
• v.30 no.2
• /
• pp.169-181
• /
• 2020

We conducted a study on the impact of observation station density; this was done in order to enable the accurate estimation of spatial meteorological variables. The purpose of this study is to help operate an efficient observation network by examining distributions of temperature, relative humidity, and wind speed in a test area of a three-dimensional meteorological observation project in the Yeongdong region in 2019. For our analysis, we grouped the observation stations as follows: 41 stations (for Step 4), 34 stations (for Step 3), 17 stations (for Step 2), and 10 stations (for Step 1). Grid values were interpolated using the kriging method. We compared the spatial accuracy of the estimated meteorological grid by using station density. The effect of increased observation network density varied and was dependent on meteorological variables and weather conditions. The temperature is sufficient for the current weather observation network (featuring an average distance about 9.30 km between stations), and the relative humidity is sufficient when the average distance between stations is about 5.04 km. However, it is recommended that all observation networks, with an average distance of approximately 4.59 km between stations, be utilized for monitoring wind speed. In addition, this also enables the operation of an effective observation network through the classification of outliers.

• Journal of Korea Water Resources Association
• /
• v.42 no.11
• /
• pp.963-972
• /
• 2009

The objective of this study is to evaluate an application of stochastic continuous storage function model with ensemble Kalman filter technique. The case study is performed at the upstream basin of Jibo streamflow gauge including Andong and Imha dam. Test period is for the rainy season during 2006 and 2007. Long term runoff analysis is feasible in the case of using deterministic model. Ensemble members for input data and parameters are generated using Monte Carlo simulation for the purpose of applying ensemble Kalman filter technique. The cumulative absolute errors of stochastic model to the deterministic one are improved for the amount of 17.5 %, 18.3 % and more than 40.0 % for Andong dam, Imha dam and Jibo station, respectively. The results indicate that the stochastic model improves the accuracy of the simulated discharge considerably.

• Atmosphere
• /
• v.16 no.2
• /
• pp.125-139
• /
• 2006

Precipitation forecasts from MM5 have been verified for the period 1989-2001 over Yeongdong region to show a tendency of model forecast. We select 57 events which are related with the heavy snowfall in Yeongdong region. They are classified into three precipitation types; mountain type, cold-coastal type, and warm type. The threat score (TS), the probability of detection (POD), and the false-alarm rate (FAR) are computed for categorical verification and the mean squared error (MSE) is also computed for scalar accuracy measures. In the case of POD, warm, mountain, and cold-coastal precipitation type are 0.71, 0.69, and 0.55 in turn, respectively. In aspect of quantitative verification, mountain and cold-coastal type are relatively well matched between forecasts and observations, while for warm type MM5 tends to overestimate precipitation. There are 12 events for the POD below 0.2, mountain, cold-coastal, warm type are 2, 7, 3 events, respectively. Most of their precipitation are distributed over the East Sea nearby Yeongdong region. These events are also shown when there are no or very weak easterlies in the lower troposphere. Even in the case that we use high resolution sea surface temperature (about 18 km) for the boundary condition, there are not much changes in the wind direction to compare that with low resolution sea surface temperature (about 100 km).

• The Journal of Information Systems
• /
• v.29 no.1
• /
• pp.241-265
• /
• 2020

Purpose Various machine learning techniques are used to implement for predicting corporate credit. However, previous research doesn't utilize time series input features and has a limited prediction timing. Furthermore, in the case of corporate bond credit rating forecast, corporate sample is limited because only large companies are selected for corporate bond credit rating. To address limitations of prior research, this study attempts to implement a predictive model with more sample companies, which can adjust the forecasting point at the present time by using the credit score information and corporate information in time series. Design/methodology/approach To implement this forecasting model, this study uses the sample of 2,191 companies with KIS credit scores for 18 years from 2000 to 2017. For improving the performance of the predictive model, various financial and non-financial features are applied as input variables in a time series through a sliding window technique. In addition, this research also tests various machine learning techniques that were traditionally used to increase the validity of analysis results, and the deep learning technique that is being actively researched of late. Findings RNN-based stateful LSTM model shows good performance in credit rating prediction. By extending the forecasting time point, we find how the performance of the predictive model changes over time and evaluate the feature groups in the short and long terms. In comparison with other studies, the results of 5 classification prediction through label reclassification show good performance relatively. In addition, about 90% accuracy is found in the bad credit forecasts.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.40 no.4
• /
• pp.253-259
• /
• 2017

Recent development in science and technology has modernized the weapon system of ROKN (Republic Of Korea Navy). Although the cost of purchasing, operating and maintaining the cutting-edge weapon systems has been increased significantly, the national defense expenditure is under a tight budget constraint. In order to maintain the availability of ships with low cost, we need accurate demand forecasts for spare parts. We attempted to find consumption pattern using data mining techniques. First we gathered a large amount of component consumption data through the DELIIS (Defense Logistics Intergrated Information System). Through data collection, we obtained 42 variables such as annual consumption quantity, ASL selection quantity, order-relase ratio. The objective variable is the quantity of spare parts purchased in f-year and MSE (Mean squared error) is used as the predictive power measure. To construct an optimal demand forecasting model, regression tree model, randomforest model, neural network model, and linear regression model were used as data mining techniques. The open software R was used for model construction. The results show that randomforest model is the best value of MSE. The important variables utilized in all models are consumption quantity, ASL selection quantity and order-release rate. The data related to the demand forecast of spare parts in the DELIIS was collected and the demand for the spare parts was estimated by using the data mining technique. Our approach shows improved performance in demand forecasting with higher accuracy then previous work. Also data mining can be used to identify variables that are related to demand forecasting.

• Journal of the Korean earth science society
• /
• v.38 no.1
• /
• pp.11-23
• /
• 2017

The short term predictability of wind resources is an important factor in evaluating the economic feasibility of a wind power plant. As a method of improving the predictability, a Bayesian Kalman filter is applied as the model data postprocessing. At this time, a statistical training period is needed to evaluate the correlation between estimated model and observation data for several Kalman training periods. This study was quantitatively analyzes for the prediction characteristics according to different training periods. The prediction of the temperature and wind speed with 3-day short term Bayesian Kalman training at Taebaek area is more reasonable than that in applying the other training periods. In contrast, it may produce a good prediction result in Ieodo when applying the training period for more than six days. The prediction performance of a Bayesian Kalman filter is clearly improved in the case in which the Weather Research Forecast (WRF) model prediction performance is poor. On the other hand, the performance improvement of the WRF prediction is weak at the accurate point.