• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.338-338
• /
• 2023

최근 특정 지역에 짧은 시간동안 많은 강우가 내리는 국지성 집중호우가 빈번히 발생하고 있으나, 이에 대한 예측과 대비에도 불구하고 피해는 지속적으로 증가하고 있다. 지속적인 강우량 증가 추이로 시간최대 및 일최대 강우량 관측기록이 해마다 갱신되고, 도시, 하천 및 주요 홍수방어 시설의 설계용량을 초과하는 피해가 발생하고 있다. 다수의 인구가 거주하고 대규모 기반시설이 집중된 도시지역에서 발생하는 집중호우는 심각한 인명 및 재산피해로 이어질 수 있다. 따라서, 부처별 재난의 저감대책은 정량적인 피해규모의 피해금액 예측보다는 설계 빈도에 대한 규모의 크기로 대책을 마련하고 있다. 국내에서는 풍수해 피해를 저감시키기 위해 개발에 따르는 재해영향요인을 개발 사업 시행 이전에 예측·분석하고 적절한 저감대책안을 수립·시행하고 있으나 설계빈도에 대한 규모일 뿐 정량적인 저감대책으로 예방되는 피해금액은 알 수 없다. 본 연구에서는 재해연보를 기반으로 호우재해(호우, 태풍)에 대한 시군구-재해기간의 피해데이터를 1999년부터 2019년까지 총 20년의 빅데이터와 전국 68개 강우관측소를 대상으로 총 20년(1999년 ~ 2019년)의 강우자료를 구축하였다. 머신러닝의 학습별 알고리즘을 조사하여 호우재해 피해데이터의 적용성이 높고 다양한 분야에 적용이 가능한 Neural networks의 분석기술인 ANN기법을 선정하였다 피해데이터의 재해발생기간별 총강우량, 일최대강우량, 총피해금액에 대하여 1999년 ~ 2018년을 학습하고 2019년에 대하여 강우특성과 피해특성의 분석하였다. 분석결과 Neural Networks의 지도학습은 총 6,902개 중 2019년을 제외한 6,414개를 학습하였으며 분석 타깃은 호우재해의 피해규모를 분석할 수 있는 총강우량, 일최대강우량, 총피해금액에 대하여 은닉노드 5개씩 2계층에 대하여 분석하였다.

• Journal of Korea Water Resources Association
• /
• v.53 no.11
• /
• pp.929-938
• /
• 2020

In this study, we estimated missing evapotranspiration (ET) data at a eddy-covariance flux tower in the Cheongmicheon farmland site using the Artificial Neural Network (ANN). The ANN showed excellent performance in numerical analysis and is expanding in various fields. To evaluate the performance the ANN-based gap-filling, ET was calculated using the existing gap-filling methods of Mean Diagnostic Variation (MDV) and Food and Aggregation Organization Penman-Monteith (FAO-PM). Then ET was evaluated by time series method and statistical analysis (coefficient of determination, index of agreement (IOA), root mean squared error (RMSE) and mean absolute error (MAE). For the validation of each gap-filling model, we used 30 minutes of data in 2015. Of the 121 missing values, the ANN method showed the best performance by supplementing 70, 53 and 84 missing values, respectively, in the order of MDV, FAO-PM, and ANN methods. Analysis of the coefficient of determination (MDV, FAO-PM, and ANN methods followed by 0.673, 0.784, and 0.841, respectively.) and the IOA (The MDV, FAO-PM, and ANN methods followed by 0.899, 0.890, and 0.951 respectively.) indicated that, all three methods were highly correlated and considered to be fully utilized, and among them, ANN models showed the highest performance and suitability. Based on this study, it could be used more appropriately in the study of gap-filling method of flux tower data using machine learning method.

• Korean Journal of Construction Engineering and Management
• /
• v.19 no.4
• /
• pp.82-92
• /
• 2018

Considering that earlier stage of construction project has a great influence on the possibility of lowering of environmental load, it is important to build and utilize system that can support effective decision making at the initial stage of the project. In this study, we constructed an environmental load estimation model that can be used at the early stage of the project using basic design factors. The model was constructed by using the artificial neural network to estimate environmental load by applying to planning stage (ANN-1), basic design stage (ANN-2). The result of test, shows that average of absolute measuring efficiency and standard deviation of ANN-1 and ANN-2 were 11.19% / 5.30% and 9.59% / 3.09% each. This result indicates that the model using the input variables extended with the project progress has high reliability and it is considered to be effective in decision support at the initial design stage of the project.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2005.05b
• /
• pp.889-893
• /
• 2005

본 연구에서는 기후변화에 따른 신뢰성 높은 수자원 영향평가를 수행하기 위한 방안으로 유출모형에 따른 불확실성을 최소화할 수 있는 슈퍼앙상블 기법을 제안하였다. 유출모형들은 자연현상을 개념화하는 과정에서 목적에 따라 알고리즘이나 구조가 다르게 개발된다. 따라서 동일한 유역에 동일한 입력자료를 사용하더라도 유출모의 결과는 상이하며 이는 곧 불확실성으로 작용한다. 이러한 불확실성을 최소화하기 위한 방법으로 본 연구에서는 통계적기법인 인공신경망 모형을 이용하여 모형별 유출결과를 향상시킬 수 있는 슈퍼앙상블 기법을 개발하고 적용성을 분석하였다. 적용 대상유역으로는 한강수계에 위치한 괴산댐유역을 선정하였으며, 적용 모형으로는 일체형 모형인 Tank 모형과 준분포형 모형인 PRMS 모형을 이용하여 슈퍼앙상블을 구축하고 검정하였다.

• Journal of Korea Water Resources Association
• /
• v.45 no.12
• /
• pp.1321-1330
• /
• 2012

In this study, the daily inflow at the basin of Chungju dam is predicted using wavelet-artificial neural network for nonlinear model. Time series generally consists of a linear combination of trend, periodicity and stochastic component. However, when framing time series model through these data, trend and periodicity component have to be removed. Wavelet transform which is denoising technique is applied to remove nonlinear dynamic noise such as trend and periodicity included in hydrometeorological data and simple noise that arises in the measurement process. The wavelet-artificial neural network (WANN) using data applied wavelet transform as input variable and the artificial neural network (ANN) using only raw data are compared. As a results, coefficient of determination and the slope through linear regression show that WANN is higher than ANN by 0.031 and 0.0115 respectively. And RMSE and RRMSE of WANN are smaller than those of ANN by 37.388 and 0.099 respectively. Therefore, WANN model applied in this study shows more accurate results than ANN and application of denoising technique through wavelet transforms is expected that more accurate predictions than the use of raw data with noise.

• 한국방재학회:학술대회논문집
• /
• 2008.02a
• /
• pp.101-104
• /
• 2008

In this study, rainfall adjust and forecasting using artificial neural network(ANN) which includes a correlation coefficient is application in Seoul region. It analyzed one-hour rainfall data which has been reported in 25 region in seoul during from 2000 to 2006 at rainfall observatory by AWS. The ANN learning algorithm apply for input data that each region using cross-correlation will use the highest correlation coefficient region. In addition, rainfall adjust analyzed the minimum error based on correlation coefficient and determination coefficient related to the input region. ANN model used back-propagation algorithm for learning algorithm. In case of the back-propagation algorithm, many attempts and efforts are required to find the optimum neural network structure as applied model. This is calculated similar to the observed rainfall that the correlation coefficient was 0.98 in missing rainfall adjust at 10 region. As a result, ANN model has been for suitable for rainfall adjust. It is considered that the result will be more accurate when it includes climate data affecting rainfall.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.125-140
• /
• 2013

We call a data set in which the number of records belonging to a certain class far outnumbers the number of records belonging to the other class, 'imbalanced data set'. Most of the classification techniques perform poorly on imbalanced data sets. When we evaluate the performance of a certain classification technique, we need to measure not only 'accuracy' but also 'sensitivity' and 'specificity'. In a customer churn prediction problem, 'retention' records account for the majority class, and 'churn' records account for the minority class. Sensitivity measures the proportion of actual retentions which are correctly identified as such. Specificity measures the proportion of churns which are correctly identified as such. The poor performance of the classification techniques on imbalanced data sets is due to the low value of specificity. Many previous researches on imbalanced data sets employed 'oversampling' technique where members of the minority class are sampled more than those of the majority class in order to make a relatively balanced data set. When a classification model is constructed using this oversampled balanced data set, specificity can be improved but sensitivity will be decreased. In this research, we developed a hybrid model of support vector machine (SVM), artificial neural network (ANN) and decision tree, that improves specificity while maintaining sensitivity. We named this hybrid model 'hybrid SVM model.' The process of construction and prediction of our hybrid SVM model is as follows. By oversampling from the original imbalanced data set, a balanced data set is prepared. SVM_I model and ANN_I model are constructed using the imbalanced data set, and SVM_B model is constructed using the balanced data set. SVM_I model is superior in sensitivity and SVM_B model is superior in specificity. For a record on which both SVM_I model and SVM_B model make the same prediction, that prediction becomes the final solution. If they make different prediction, the final solution is determined by the discrimination rules obtained by ANN and decision tree. For a record on which SVM_I model and SVM_B model make different predictions, a decision tree model is constructed using ANN_I output value as input and actual retention or churn as target. We obtained the following two discrimination rules: 'IF ANN_I output value <0.285, THEN Final Solution = Retention' and 'IF ANN_I output value ${\geq}0.285$, THEN Final Solution = Churn.' The threshold 0.285 is the value optimized for the data used in this research. The result we present in this research is the structure or framework of our hybrid SVM model, not a specific threshold value such as 0.285. Therefore, the threshold value in the above discrimination rules can be changed to any value depending on the data. In order to evaluate the performance of our hybrid SVM model, we used the 'churn data set' in UCI Machine Learning Repository, that consists of 85% retention customers and 15% churn customers. Accuracy of the hybrid SVM model is 91.08% that is better than that of SVM_I model or SVM_B model. The points worth noticing here are its sensitivity, 95.02%, and specificity, 69.24%. The sensitivity of SVM_I model is 94.65%, and the specificity of SVM_B model is 67.00%. Therefore the hybrid SVM model developed in this research improves the specificity of SVM_B model while maintaining the sensitivity of SVM_I model.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.6
• /
• pp.77-82
• /
• 2005

This paper is proposed fuzzy neural network(FNN) and artificial neural network(ANN) based on the vector controlled induction motor drive system. The hybrid combination of fuzzy control and neural network will produce a powerful representation flexibility and numerical processing capability. Also, this paper is proposed control and estimation of speed of induction motor using fuzzy and neural network. The back propagation neural network technique is used to provide a real time adaptive estimation of the motor speed. The error between the desired state variable and the actual one is back-propagated to adjust the rotor speed, so that the actual state variable will coincide with the desired one. The back propagation mechanism is easy to derive and the estimated speed tracks precisely the actual motor speed. This paper is proposed the experimental results to verify the effectiveness of the new method.

• Journal of Korea Water Resources Association
• /
• v.49 no.10
• /
• pp.877-885
• /
• 2016

In this study, the Standard Precipitation Index(SPI), meteorological drought index, was used to evaluate the temporal and spatial assessment of drought forecasting results for all cross Korea. For the drought forecasting, the Multi Layer Perceptron-Artificial Neural Network (MLP-ANN) was selected and the drought forecasting was performed according to different forecasting lead time for SPI (3) and SPI (6). The precipitation data observed in 59 gaging stations of Korea Meteorological Adminstration (KMA) from 1976~2015. For the performance evaluation of the drought forecasting, the binary classification confusion matrix, such as evaluating the status of drought occurrence based on threshold, was constituted. Then Receiver Operating Characteristics (ROC) score and F score according to conditional probability are computed. As a result of ROC analysis on forecasting performance, drought forecasting performance, of applying the MLP-ANN model, shows satisfactory forecasting results. Consequently, two-month and five-month leading forecasts were possible for SPI (3) and SPI (6), respectively.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.1-4
• /
• 2001

The classification performance of Artificial Neural Network (ANN) and RBF-NN was compared for Landsat TM image. The RBF-NN was validated for three unique landuse types (e.g. Mixed landuse area, Cultivated area, Urban area), different input band combinations and classification class. The bootstrap resampling technique was employed to estimate the confidence intervals and distribution for unit load, The pollutant generation was varied significantly according to the classification accuracy and percentile unit load applied. Especially in urban area, where mixed landuse is dominant, the difference of estimated pollutant load is largely varied.