• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.112-112
• /
• 2020

하천에서 유해화학물질 유입 사고 발생 시 수환경 피해를 최소화하기 위해 신속한 초기 대응이 필요하다. 따라서, 본 연구에서는 수환경 화학사고 대응 시스템 구축을 위해 하천 실시간 모니터링 지점에서 관측된 유해화학물질의 농도 자료를 이용하여 발생원의 유입 지점과 유입량을 역추적하는 프레임워크를 개발하였다. 본 연구에서 제시하는 프레임워크는 첫 번째로 하천 저장대 모형(Transient Storage Zone Model; TSM)과 HEC-RAS 모형을 이용하여 다양한 유량의 수리 조건에서 화학사고 시나리오를 생성하는 단계, 두번째로 생성된 시나리오의 유입 지점과 유입량에 대한 시간-농도 곡선 (BreakThrough Curve; BTC)을 21개의 곡선특징 (BTC feature)으로 추출하는 단계, 최종적으로 재귀적 특징 선택법(Recursive Feature Elimination; RFE)을 이용하여 의사결정나무 모형, 랜덤포레스트 모형, Xgboost 모형, 선형 서포트 벡터 머신, 커널 서포트 벡터 머신 그리고 Ridge 모형에 대한 모형별 주요 특징을 학습하고 성능을 비교하여 각각 유입 위치와 유입 질량 예측에 대한 최적 모형 및 특징 조합을 제시하는 단계로 구축하였다. 또한, 현장 적용성 제고를 위해 시간-농도 곡선을 2가지 경우 (Whole BTC와 Fractured BTC)로 가정하여 기계학습 모형을 학습시켜 모의결과를 비교하였다. 제시된 프레임워크의 검증을 위해서 낙동강 지류인 감천에 적용하여 모형을 구축하고 시나리오 자료 기반 검증과 Rhodamine WT를 이용한 추적자 실험자료를 이용한 검증을 수행하였다. 기계학습 모형들의 비교 검증 결과, 각 모형은 가중항 기반과 불순도 감소량 기반 특징 중요도 산출 방식에 따라 주요 특징이 상이하게 산출되었으며, 전체 시간-농도 곡선 (WBTC)과 부분 시간-농도 곡선 (FBTC)별 최적 모형도 다르게 산출되었다. 유입 위치 정확도 및 유입 질량 예측에 대한 R²는 대부분의 모형이 90% 이상의 우수한 결과를 나타냈다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.136-136
• /
• 2022

This study investigates the possibility of coupling empirical mode decomposition (EMD) for runoff prediction from machine learning (ML) models. Here, support vector regression (SVR) and convolutional neural network (CNN) were considered for ML algorithms. Precipitation (P), minimum temperature (Tmin), maximum temperature (Tmax) and their intrinsic mode functions (IMF) values were used for input variables at a monthly scale from Jan. 1973 to Dec. 2020 in the Grand river basin, Canada. The support vector machine-recursive feature elimination (SVM-RFE) technique was applied for finding the best combination of predictors among input variables. The results show that the proposed method outperformed the individual performance of SVR and CNN during the training and testing periods in the study area. According to the correlation coefficient (R), the EMD-SVR model outperformed the EMD-CNN model in both training and testing even though the CNN indicated a better performance than the SVR before using IMF values. The EMD-SVR model showed higher improvement in R value (38.7%) than that from the EMD-CNN model (7.1%). It should be noted that the coupled models of EMD-SVR and EMD-CNN represented much higher accuracy in runoff prediction with respect to the considered evaluation indicators, including root mean square error (RMSE) and R values.

• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.173-198
• /
• 2020

For a long time, many studies have been conducted on predicting the success of campaigns for customers in academia, and prediction models applying various techniques are still being studied. Recently, as campaign channels have been expanded in various ways due to the rapid revitalization of online, various types of campaigns are being carried out by companies at a level that cannot be compared to the past. However, customers tend to perceive it as spam as the fatigue of campaigns due to duplicate exposure increases. Also, from a corporate standpoint, there is a problem that the effectiveness of the campaign itself is decreasing, such as increasing the cost of investing in the campaign, which leads to the low actual campaign success rate. Accordingly, various studies are ongoing to improve the effectiveness of the campaign in practice. This campaign system has the ultimate purpose to increase the success rate of various campaigns by collecting and analyzing various data related to customers and using them for campaigns. In particular, recent attempts to make various predictions related to the response of campaigns using machine learning have been made. It is very important to select appropriate features due to the various features of campaign data. If all of the input data are used in the process of classifying a large amount of data, it takes a lot of learning time as the classification class expands, so the minimum input data set must be extracted and used from the entire data. In addition, when a trained model is generated by using too many features, prediction accuracy may be degraded due to overfitting or correlation between features. Therefore, in order to improve accuracy, a feature selection technique that removes features close to noise should be applied, and feature selection is a necessary process in order to analyze a high-dimensional data set. Among the greedy algorithms, SFS (Sequential Forward Selection), SBS (Sequential Backward Selection), SFFS (Sequential Floating Forward Selection), etc. are widely used as traditional feature selection techniques. It is also true that if there are many risks and many features, there is a limitation in that the performance for classification prediction is poor and it takes a lot of learning time. Therefore, in this study, we propose an improved feature selection algorithm to enhance the effectiveness of the existing campaign. The purpose of this study is to improve the existing SFFS sequential method in the process of searching for feature subsets that are the basis for improving machine learning model performance using statistical characteristics of the data to be processed in the campaign system. Through this, features that have a lot of influence on performance are first derived, features that have a negative effect are removed, and then the sequential method is applied to increase the efficiency for search performance and to apply an improved algorithm to enable generalized prediction. Through this, it was confirmed that the proposed model showed better search and prediction performance than the traditional greed algorithm. Compared with the original data set, greed algorithm, genetic algorithm (GA), and recursive feature elimination (RFE), the campaign success prediction was higher. In addition, when performing campaign success prediction, the improved feature selection algorithm was found to be helpful in analyzing and interpreting the prediction results by providing the importance of the derived features. This is important features such as age, customer rating, and sales, which were previously known statistically. Unlike the previous campaign planners, features such as the combined product name, average 3-month data consumption rate, and the last 3-month wireless data usage were unexpectedly selected as important features for the campaign response, which they rarely used to select campaign targets. It was confirmed that base attributes can also be very important features depending on the type of campaign. Through this, it is possible to analyze and understand the important characteristics of each campaign type.

• Journal of Korean Society of Rural Planning
• /
• v.28 no.1
• /
• pp.57-69
• /
• 2022

Crop classification is very important for estimating crop yield and figuring out accurate cultivation area. The purpose of this study is to classify crops harvested in fall in Idam-ri, Goesan-gun, Chungcheongbuk-do by using unmanned aerial vehicle (UAV) images and support vector machine (SVM) model. The study proceeded in the order of image acquisition, variable extraction, model building, and evaluation. First, RGB and multispectral image were acquired on September 13, 2021. Independent variables which were applied to Farm-Map, consisted gray level co-occurrence matrix (GLCM)-based texture characteristics by using RGB images, and multispectral reflectance data. The crop classification model was built using texture characteristics and reflectance data, and finally, accuracy evaluation was performed using the error matrix. As a result of the study, the classification model consisted of four types to compare the classification accuracy according to the combination of independent variables. The result of four types of model analysis, recursive feature elimination (RFE) model showed the highest accuracy with an overall accuracy (OA) of 88.64%, Kappa coefficient of 0.84. UAV-based RGB and multispectral images effectively classified cabbage, rice and soybean when the SVM model was applied. The results of this study provided capacity usefully in classifying crops using single-period images. These technologies are expected to improve the accuracy and efficiency of crop cultivation area surveys by supplementing additional data learning, and to provide basic data for estimating crop yields.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.413-413
• /
• 2019

하천과 호수에서 남조류의 이상 과잉증식 문제(이하 녹조문제)는 담수생태계의 생물다양성을 감소시키며, 음용수의 이취미 원인물질을 발생시켜 물 이용에 장해가 된다. 또한 독소를 생산하는 유해남조류가 대량 증식할 경우에는 가축이나 인간의 건강에 치명적 해를 끼치기도 한다. 그 동안 국내에서 녹조문제는 댐 저수지와 하구호와 같은 정체수역에서 간헐적으로 문제를 일으켰으나, 4대강사업(2010-2011)으로 16개의 보가 설치된 이후 낙동강, 금강, 영산강 등 대하천에서도 광범위하게 발생되고 있어 중요한 사회적 환경적 이슈로 대두되었다. 한편, 대하천에 설치된 보 구간에서 빈번히 발생하는 녹조현상의 원인에 대해서는 전 지구적 기온상승에 따른 기후변화의 영향이라는 주장과 유역으로부터 영양염류의 과도한 유입, 가뭄에 따른 유량감소, 보 설치에 따른 체류시간 증가 등 다양한 의견이 제시되고 있으나, 대상 유역과 수체의 특성에 따라 녹조 발생의 원인이 상이하거나 또는 다양한 요인이 복합적으로 작용하기 때문에 보편적 해석(universal interpretation)이 어려운 것이 현실이다. 따라서 각 수계별, 보별 녹조현상에 대한 정확한 원인분석과 효과적인 대책 마련을 위해서는 집중된 실험자료와 데이터마이닝 기법에 근거로 한 보다 과학적이고 객관적인 접근이 이루어져야 한다. 본 연구에서는 2012년 보 설치 이후 남조류에 의한 녹조현상이 빈번히 발생하고 있는 낙동강 4개보(강정고령보, 달성보, 합천창녕보, 창녕함안보)를 대상으로 집중적인 현장조사와 실험분석을 수행하고, 수집된 기상, 수문, 수질, 조류 자료에 대해 통계분석과 다양한 데이터모델링 기법을 적용하여 보별 남조류 우점 환경조건과 이를 제어하기 위한 주요 조절변수를 규명하는데 있다. 연구대상 보 별 수질과 식물플랑크톤의 정성 및 정량 실험은 2017년 5월부터 2018년 11월까지 2년에 걸쳐 실시하였으며, 남조류 세포수 밀도와 환경요인과의 상관성 분석을 실시하고, 단계적 다중회귀모델(Step-wise Multiple Linear Regressions, SMLR), 랜덤포레스트(Random Forests, RF) 모델과 재귀적 변수 제거 기법(Recursive Feature Elimination using Random Forest, RFE-RF)을 이용한 변수중요도 평가, 의사결정나무(Decision Tree, DT), 주성분분석(Principal Component Analysis, PCA) 기법 등 다양한 모수적 및 비모수적 데이터마이닝 결과를 바탕으로 각 보별 남 조류 우점 환경요인을 종합적으로 해석하였다.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.43 no.3
• /
• pp.321-335
• /
• 2023

Monitoring sediment loads in natural rivers is the key process in river engineering, but it is costly and dangerous. In practice, suspended loads are directly measured, and total loads, which is a summation of suspended loads and bed loads, are estimated. This study proposes a real-time sediment discharge monitoring system using the horizontal acoustic Doppler current profiler (H-ADCP) and support vector regression (SVR). The proposed system is comprised of the SVR model for suspended sediment concentration (SVR-SSC) and for total loads (SVR-Q_TL), respectively. SVR-SSC estimates SSC and SVR-Q_TL mimics the modified Einstein procedure. The grid search with K-fold cross validation (Grid-CV) and the recursive feature elimination (RFE) were employed to determine SVR's hyperparameters and input variables. The two SVR models showed reasonable cross-validation scores (R²) with 0.885 (SVR-SSC) and 0.860 (SVR-Q_TL). During the time-series sediment load monitoring period, we successfully detected various sediment transport phenomena in natural streams, such as hysteresis loops and sensitive sediment fluctuations. The newly proposed sediment monitoring system depends only on the gauged features by H-ADCP without additional assumptions in hydraulic variables (e.g., friction slope and suspended sediment size distribution). This method can be applied to any ADCP-installed discharge monitoring station economically and is expected to enhance temporal resolution in sediment monitoring.