• Proceedings of the Korea Water Resources Association Conference
• /
• 2002.05a
• /
• pp.59-64
• /
• 2002

An expert system, to assist dam managers for five dams along the Saikawa River, has been developed with a primary objective of achieving swift and accurate reservoir operation decision-makings during floods. The expert system is capable of supporting on decision-makings upon establishment of flood management procedure and release/storage planning. Furthermore, an attempt was made to improve reservoir inflow prediction models for better supporting capability. As a result, accuracy on prediction of inflow up to 7 hours ahead was improved, which is important for flood management of the five dams, using neural network. The neural network inflow prediction models were developed for each types of floods caused by frontal rainfalls, snowmelt and typhoons, after extracting relevant meteorological factors for each.

• The Journal of the Korea Contents Association
• /
• v.20 no.10
• /
• pp.702-714
• /
• 2020

Stock investment is a personal investment technique that has gathered tremendous interest since the reduction in interest rates and tax exemption. However, it is risky especially for those who do not have expert knowledge on stock volatility. Therefore, it is well understood that accurate stock trend prediction can greatly help stock investment, giving birth to a volume of research work in the field. In order to compare different research works and to optimize hyper-parameters for prediction models, it is required to have an evaluation standard that can accurately assess performances of prediction models. However, little research has been done in the area, and conventionally used methods have been employed repeatedly without being rigorously validated. For this reason, we first analyze performance evaluation of stock trend prediction with respect to performance metrics and data composition, and propose a fair evaluation method based on prediction disparity ratio.

• The Journal of the Korea Contents Association
• /
• v.16 no.11
• /
• pp.488-498
• /
• 2016

Much of the information is stored as data, research has been activated for analyzing the data and predicting the special circumstances. In the case of power data, the studies, such as research of renewable energy utilization, power prediction depending on site characteristics, smart grid, and micro-grid, is actively in progress. In this paper, we propose a power prediction model using the substation environment data. In this case, we try to verify the power prediction result to reflect the multiple arguments on the power and weather data, rather than a simple power data. The validation process is the effect of multiple factors compared to other two methods, one of power prediction result considering power data and the other result using power pattern data that have been made in the similar weather data. Our system shows that it can achieve max prediction error of less than 15%.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.281-281
• /
• 2015

Shallow landslide often occurs in areas of this topography where subsurface soil water flow paths give rise to excess pore-water pressures downslope. Recent hillslope hydrology studies have shown that subsurface topography has a strong impact in controlling the connectivity of saturated areas at the soil-bedrock interface. In this study, the physically based SHALSTAB model was used to evaluate the effects of three soil thicknesses (i.e. average soil layer, soil thickness to weathered soil and soil thickness to bedrock soil layer) and subsurface flow reflecting three soil thicknesses on shallow landslide prediction accuracy. Three digital elevation models (DEMs; i.e. ground surface, weathered surface and bedrock surface) and three soil thicknesses (average soil thickness, soil thickness to weathered rock and soil thickness to bedrock) at a small hillslope site in Jinbu, Kangwon Prefecture, eastern part of the Korean Peninsula, were considered. Each prediction result simulated with the SHALSTAB model was evaluated by receiver operating characteristic (ROC) analysis for modelling accuracy. The results of the ROC analysis for shallow landslide prediction using the ground surface DEM (GSTO), the weathered surface DEM and the bedrock surface DEM (BSTO) indicated that the prediction accuracy was higher using flow accumulation by the BSTO and weathered soil thickness compared to results. These results imply that 1) the effect of subsurface flow by BSTO on shallow landslide prediction especially could be larger than the effects of topography by GSTO, and 2) the effect of weathered soil thickness could be larger than the effects of average soil thickness and bedrock soil thickness on shallow landslide prediction. Therefore, we suggest that using BSTO dem and weathered soil layer can improve the accuracy of shallow landslide prediction, which should contribute to more accurately predicting shallow landslides.

• Korean Journal of Hydrosciences
• /
• v.6
• /
• pp.1-11
• /
• 1995

A plunge pool is often employed as an energy-dissipating device at the end of a spillway or a pipe culvert. A jet from spillways or pipes frequently generates a scour hole which threaten the stability of the hydraulic structure. Existing scour prediction formulas of plunge pool of spillways or pipe culverts give a wide range of scour depths, and it is, therefore, difficult to accurately predict those scour depths. In this study, a new experimental method and new sour prediction formulas under submerged circular jet for large bed materials with shallow tailwater depths were developed. A major variable, which was not used in previous scour prediction equations, was the ratio of jet size to bed material size. In this study, jet momentum acting on a bed particle and jet diffustion theory were employed to derive scour prediction formulas. Four theoretical formulas were suggested for the two regions of jet diffusion, i.e., the region of flow establishment and the region of established flow. The semi-theoretically developed scour prediction formulas showed close agreement with laboratory experiments performed on movable bed made of large spherical particles.

• Environmental and Resource Economics Review
• /
• v.11 no.3
• /
• pp.447-463
• /
• 2002

The heat demand prediction is an essential issue in management of district heating system. Without an accurate prediction through the lead-time period, it might be impossible to make a rational decision on many issues such as heat production scheduling and heat exchange among the plants which are very critical for the district heating company. The heat demand varies with the temperature as well as the time nonlinearly. And the parametric specification of the heat demand model would cause a misspecification bias in prediction. A nonparametric model for the short-term heat demand prediction has been developed as an alternative to avoiding the misspecification error and tested with the actual data. The prediction errors are reasonably small enough to use the model to predict a few hour ahead heat demand.

• Journal of Korea Water Resources Association
• /
• v.54 no.5
• /
• pp.301-309
• /
• 2021

In this study, we tried to improve the performance of the existing U-net-based deep learning rainfall prediction model, which can weaken the meaning of time series order. For this, ConvLSTM2D U-Net structure model considering temporal consistency of data was applied, and we evaluated accuracy of the ConvLSTM2D U-Net model using a RainNet model and an extrapolation-based advection model. In addition, we tried to improve the uncertainty in the model training process by performing learning not only with a single model but also with 10 ensemble models. The trained neural network rainfall prediction model was optimized to generate 10-minute advance prediction data using four consecutive data of the past 30 minutes from the present. The results of deep learning rainfall prediction models are difficult to identify schematically distinct differences, but with ConvLSTM2D U-Net, the magnitude of the prediction error is the smallest and the location of rainfall is relatively accurate. In particular, the ensemble ConvLSTM2D U-Net showed high CSI, low MAE, and a narrow error range, and predicted rainfall more accurately and stable prediction performance than other models. However, the prediction performance for a specific point was very low compared to the prediction performance for the entire area, and the deep learning rainfall prediction model also had limitations. Through this study, it was confirmed that the ConvLSTM2D U-Net neural network structure to account for the change of time could increase the prediction accuracy, but there is still a limitation of the convolution deep neural network model due to spatial smoothing in the strong rainfall region or detailed rainfall prediction.

• Asian-Australasian Journal of Animal Sciences
• /
• v.28 no.7
• /
• pp.993-998
• /
• 2015

Possibilities of predicting eggshell ultrastructure from direct non-destructive and destructive measurements were examined using 120 Fayoumi eggs collected from the flock at 45 weeks of age. The non-destructive measurements included weight, length and width of the egg. The destructive measurements were breaking strength and shell thickness. The eggshell ultrastructure traits involved the total thickness of eggshell layer, thickness of palisade layer, cone layer and total score. Prediction of total thickness of eggshell layer based on non-destructive measurements individually or simultaneously was not possible ($R^2=0.01$ to 0.16). The destructive measurements were far more accurate than the non-destructive in predicting total thickness of eggshell layer. Prediction based on breaking strength alone was more accurate ($R^2=0.85$) than that based on shell thickness alone ($R^2=0.72$). Adding shell thickness to breaking strength (the best predictor) increased the accuracy of prediction by 5%. The results obtained indicated that both non-destructive and destructive measurements were not useful in predicting the cone layer ($R^2$ not exceeded 18%). The maximum accuracy of prediction of total score ($R^2=0.48$) was obtained from prediction based on breaking strength alone. Combining shell thicknesses and breaking strength into one equation was no help in improving the accuracy of prediction.

• The Journal of the Korea Contents Association
• /
• v.10 no.6
• /
• pp.134-143
• /
• 2010

This paper proposes a framework for classifying metric-based software quality prediction models, especially case of software criticality, into four types. Models are classified along two vectors: input metric forms and the necessity of past project data. Each type has its own characteristics and its strength and weakness are compared with those of other types using newly defined criteria. Through this qualitative evaluation each organization can choose a proper model to suit its environment. My earlier studies of criticality prediction model implemented specific models in each type and evaluated their prediction performances. In this paper I analyze the experimental results and show that the characteristics of a model type is the another key of successful model selection.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.3 no.3
• /
• pp.45-56
• /
• 2001

It is very important to survey in construction for acquiring informations about geological structure which influence the tunnel construction. Usually, TSP(Tunnel Seismic Prediction) one of the seismic survey method in tunnels in Korea has been used for prediction of geological condition ahead of tunnel face, but in this study HSP(Horizontal Seismic Profiling) was used. As a result of field application, we predicted that there exist rock discontinuity such as fault, joints and bedding planes. In addition, RMR value from rock mass evaluation coincided with that from seismic survey for rock condition. We compared with rock classifications and evaluation results for proving.