• Journal of Korea Water Resources Association
• /
• v.38 no.12 s.161
• /
• pp.1051-1060
• /
• 2005

The present study analyzes hydrologic utilization of optimal radar-derived rainfall by using semi-distributed TOPMODEL and evaluates the impacts of radar rainfall and model parametric uncertainty on a hydrologic model. Monte Carlo technique is used to produce the flow ensembles. The simulated flows from the corrected radar rainfalls with real-time bias adjustment scheme are well agreed to observed flows during 22-26 July 2003. It is shown that radar-derived rainfall is useful for simulating streamflow on a basin scale. These results are diagnose with which radar-rainfall Input and parametric uncertainty influence the character of the flow simulation uncertainty. The main conclusions for this uncertainty analysis are that the radar input uncertainty is less influent than the parametric one, and combined uncertainty with radar and Parametric input can be included the highest uncertainty on a streamflow simulation.

• Atmosphere
• /
• v.23 no.1
• /
• pp.113-121
• /
• 2013

The objective of this study is to develop and evaluate the system to produce the real-time ensemble drought prediction data. Ensemble drought prediction consists of 3 processes (meteorological outlook using the multi-initial conditions, hydrological analysis and drought index calculation) therefore, more processing time and data is required than that of single member. For ensemble drought prediction, data process time is optimized and hardware of existing system is upgraded. Ensemble drought data is estimated for year 2012 and to evaluate the accuracy of drought prediction data by using ROC (Relative Operating Characteristics) analysis. We obtained 5 ensembles as optimal number and predicted drought condition for every tenth day i.e. 5th, 15th and 25th of each month. The drought indices used are SPI (Standard Precipitation Index), SRI (Standard Runoff Index), SSI (Standard Soil moisture Index). Drought conditions were determined based on results obtained for each ensemble member. Overall the results showed higher accuracy using ensemble members as compared to single. The ROC score of SRI and SSI showed significant improvement in drought period however SPI was higher in the demise period. The proposed ensemble drought prediction system can be contributed to drought forecasting techniques in Korea.

• Computers and Concrete
• /
• v.8 no.6
• /
• pp.647-675
• /
• 2011

Our ability to predict hydration behavior is becoming increasingly relevant to the concrete community as modelers begin to link material performance to the dynamics of material properties and chemistry. At early ages, the properties of concrete are changing rapidly due to chemical transformations that affect mechanical, thermal and transport responses of the composite. At later ages, the resulting, nano-, micro-, meso- and macroscopic structure generated by hydration will control the life-cycle performance of the material in the field. Ultimately, creep, shrinkage, chemical and physical durability, and all manner of mechanical response are linked to hydration. As a way to enable the modeling community to better understand hydration, a review of hydration models is presented offering insights into their mathematical origins and relationships one-to-the-other. The quest for a universal model begins in the 1920's and continues to the present, and is marked by a number of critical milestones. Unfortunately, the origins and physical interpretation of many of the most commonly used models have been lost in their overuse and the trail of citations that vaguely lead to the original manuscripts. To help restore some organization, models were sorted into four categories based primarily on their mathematical and theoretical basis: (1) mass continuity-based, (2) nucleation-based, (3) particle ensembles, and (4) complex multi-physical and simulation environments. This review provides a concise catalogue of models and in most cases enough detail to derive their mathematical form. Furthermore, classes of models are unified by linking them to their theoretical origins, thereby making their derivations and physical interpretations more transparent. Models are also used to fit experimental data so that their characteristics and ability to predict hydration calorimetry curves can be compared. A sort of evolutionary tree showing the progression of models is given along with some insights into the nature of future work yet needed to develop the next generation of cement hydration models.

• Journal of KIISE:Software and Applications
• /
• v.30 no.3_4
• /
• pp.251-258
• /
• 2003

A support vector machine (SVM) is supposed to provide a good generalization performance, but the actual performance of a actually implemented SVM is often far from the theoretically expected level. This is largely because the implementation is based on an approximated algorithm, due to the high complexity of time and space. To improve this limitation, we propose ensemble of SVMs by using Bagging (bootstrap aggregating) and Boosting. By a Bagging stage each individual SVM is trained independently using randomly chosen training samples via a bootstrap technique. By a Boosting stage an individual SVM is trained by choosing training samples according to their probability distribution. The probability distribution is updated by the error of independent classifiers, and the process is iterated. After the training stage, they are aggregated to make a collective decision in several ways, such ai majority voting, the LSE(least squares estimation) -based weighting, and double layer hierarchical combining. The simulation results for IRIS data classification, the hand-written digit recognition and Face detection show that the proposed SVM ensembles greatly outperforms a single SVM in terms of classification accuracy.

• The Korean Journal of Applied Statistics
• /
• v.27 no.3
• /
• pp.357-371
• /
• 2014

There are many studies related to imbalanced data in which the class distribution is highly skewed. To address the problem of imbalanced data, previous studies deal with resampling techniques which correct the skewness of the class distribution in each sampled subset by using under-sampling, over-sampling or hybrid-sampling such as SMOTE. Ensemble methods have also alleviated the problem of class imbalanced data. In this paper, we compare around a dozen algorithms that combine the ensemble methods and resampling techniques based on simulated data sets generated by the Backbone model, which can handle the imbalance rate. The results on various real imbalanced data sets are also presented to compare the effectiveness of algorithms. As a result, we highly recommend the resampling technique combining ensemble methods for imbalanced data in which the proportion of the minority class is less than 10%. We also find that each ensemble method has a well-matched sampling technique. The algorithms which combine bagging or random forest ensembles with random undersampling tend to perform well; however, the boosting ensemble appears to perform better with over-sampling. All ensemble methods combined with SMOTE outperform in most situations.

• Journal of Convergence for Information Technology
• /
• v.8 no.1
• /
• pp.201-206
• /
• 2018

Recently, as technologies for realizing artificial intelligence have become more common, machine learning is widely used. Machine learning provides insight into collecting large amounts of data, batch processing, and taking final action, but the effects of the work are not immediately integrated into the learning process. In this paper proposed an adaptive learning model to improve the performance of real-time stream analysis as a big business issue. Adaptive learning generates the ensemble by adapting to the complexity of the data set, and the algorithm uses the data needed to determine the optimal data point to sample. In an experiment for six standard data sets, the adaptive learning model outperformed the simple machine learning model for classification at the learning time and accuracy. In particular, the support vector machine showed excellent performance at the end of all ensembles. Adaptive learning is expected to be applicable to a wide range of problems that need to be adaptively updated in the inference of changes in various parameters over time.

• Journal of the Korean Vacuum Society
• /
• v.15 no.3
• /
• pp.308-313
• /
• 2006

We report a photoluminescence (PL) study on the growth process of self-assembled InAs quantum dots (QDs) under the various growth conditions. Distinctive double-peak feature was observed in the PL spectra of the QD samples grown at the relatively high substrate temperature. From the excitation power-dependent PL and the temperature-dependent PL measurements, the double-peak feature is associated with the ground state transitions from InAs QDs with two different size branches. In addition, the variation in the bimodal size distribution of the QD ensembles with different InAs coverage is demonstrated.

• Applied Chemistry for Engineering
• /
• v.3 no.1
• /
• pp.24-34
• /
• 1992

Recent studies dealing with the fundamental understanding and applications of bimetallic catalysts are discussed. Bimetallic catalysts have had a major industrial impact, specifically for the reforming of petroleum naphtha, for the hydrogen reduction of carbon monoxide, and for the three way catalytic converter system. The action of the bimetallic catalysts in these reactions may be interpreted in terms of ensembles, electronic influences and surface structure. Various combinations of metal pairs have been considered in order to evaluate the role played by the added metals. For catalyst selectivity control, the possibility of surface enrichment of one element has been recognised. More generally, the influence of preparative variables on the formation of supported catalysts has been clarified, In particular by temperature programmed reduction (TPR). Information on the structure of bimetallic catalysts has been obtained with chemical probes, such as chemisorption and reaction rate measurement and physical probes, such as extended X-ray absorption fine structure (EXAFS), scanning transmission electron microscopy (STEM) and Xe-NMR.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.4
• /
• pp.992-999
• /
• 2014

This paper presents an extensive experimental comparison of a variety of multi-label learning methods for the accurate prediction of subcellular localization of proteins which simultaneously exist at multiple subcellular locations. We compared several methods from three categories of multi-label classification algorithms: algorithm adaptation, problem transformation, and meta learning. Experimental results are analyzed using 12 multi-label evaluation measures to assess the behavior of the methods from a variety of view-points. We also use a new summarization measure to find the best performing method. Experimental results show that the best performing methods are power-set method pruning a infrequently occurring subsets of labels and classifier chains modeling relevant labels with an additional feature. futhermore, ensembles of many classifiers of these methods enhance the performance further. The recommendation from this study is that the correlation of subcellular locations is an effective clue for classification, this is because the subcellular locations of proteins performing certain biological function are not independent but correlated.

• Journal of Korea Water Resources Association
• /
• v.46 no.3
• /
• pp.229-243
• /
• 2013

Hydrological simulation Program-Fortran (HSPF) model was used to simulate streamflow and snow depth at Pyengchang watershed. The selected Global Climate Models (GCMs) provided by the Coupled Model Intercomparision Project Phase 3 (CMIP3) were utilized to evaluate streamflow and snow depth driven by future climate scenarios, including A1, A1B, and B1. Bias-correlation and temporal downscaling processes have been performed to minimize systematic errors between GCMs and HSPF. Based on simulated monthly streamflow and snow depth after calibration, the results indicate that HSPF performs well. The correlation coefficient between the observed and simulated monthly streamflow is 0.94. Snow depth simulations also show high correlation coefficient, which is 0.91. The results indicate that snow depth in 2018 at Pyongchang winter olympic venues will decrease by 17.62%, 9.38%, and 7.25% in January, February, and March respectively, based on streamflow realizations induced by all GCMs ensembles.