Ensemble Model Output Statistics를 이용한 평창지역 다중 모델 앙상블 결합 및 보정

황유선(Yuseon Hwang),김찬수(Chansoo Kim) 한국기상학회 2018 대기 Vol.28 No.3

The objective of this paper is to compare probabilistic temperature forecasts from different regional and global ensemble prediction systems over PyeongChang area. A statistical post-processing method is used to take into account combination and calibration of forecasts from different numerical prediction systems, laying greater weight on ensemble model that exhibits the best performance. Observations for temperature were obtained from the 30 stations in PyeongChang and three different ensemble forecasts derived from the European Centre for Medium-Range Weather Forecasts, Ensemble Prediction System for Global and Limited Area Ensemble Prediction System that were obtained between 1 May 2014 and 18 March 2017. Prior to applying to the post-processing methods, reliability analysis was conducted to identify the statistical consistency of ensemble forecasts and corresponding observations. Then, ensemble model output statistics and bias-corrected methods were applied to each raw ensemble model and then proposed weighted combination of ensembles. The results showed that the proposed methods provide improved performances than raw ensemble mean. In particular, multi-model forecast based on ensemble model output statistics was superior to the bias-corrected forecast in terms of deterministic prediction.

임의변수선택 기반 앙상블 판별분석에서 변수의 상대적 중요도에 관한 연구

최영득,임경덕,임대환,조영래,구자용 한국자료분석학회 2014 Journal of the Korean Data Analysis Society Vol.16 No.6

A method for enhancing stability and precision of a classification method is ensemble method. Typically, ensemble method outperforms a single classifier in the binary classification. The prediction structure of the ensemble model is more complex than a single base learner so that it is difficult to identify the role and importance of the explanatory variables. Ensemble method has a problem in interpretation of the prediction result since the interpretability of the prediction result of an ensemble method may be reduced. Hence, it is hard to achieve both the performance improvement in the precision and the result understanding in the interpretation of the model because of being many variables in classification model. This paper considers a methodology to solve the interpretation problem for an ensemble method. First, we explain ensemble method using random predictor selection which combines logistic regression model, a technique being used in credit scoring. Next, as a measure for the relative importance, we adopt the mean z-score that measures role and relative importance of the explanatory variables and examine the interpretation ability of the prediction result. In order to illustrate the finite-sample performance of the considered methodology, we conduct a numerical study using both a simulated data set and real data set. 이범주 판별문제에서 단일판별기보다 모형의 안정성 및 정밀도를 높이기 위한 방법론 중의 하나가 앙상블 기법이다. 앙상블 기법을 적용하는 경우에는 모형의 예측구조가 복잡하여 각 설명변수의 역할과 중요성을 확인하기 어려워 결국에는 예측결과의 해석력이 떨어진다는 단점을 가지고 있다. 본 논문에서는 임의변수선택 기반 앙상블 판별분석에서 변수의 상대적 중요도를 연구하고자 한다. 먼저 신용평점화에서 핵심기법으로 사용되는 로직스틱 회귀모형을 결합하는 임의변수선택을 이용한 앙상블 기법을 설명하고, 다음으로 각 설명변수의 역할과 상대적 중요도를 측정할 수 있는 평균 z-스코어를 이용하는 방법을 제안하여 예측결과에 대한 해석력을 살펴보고자 한다. 본 논문에서 제안한 방법론의 유한표본 성질을 규명하고 응용성을 확인하기 위하여 모의실험과 실제자료 분석을 이용한 연구를 수행하고자 한다.

평창 지역 기상예측에 대한 다중모델 앙상블의 보정

김찬수 건국대학교 기후연구소 2017 기후연구 Vol.12 No.4

In this study, a weighted ensemble method of numerical weather prediction by ensemble models is applied for PyeongChang area. The post-processing method takes into account combination and calibration of forecasts from different numerical models, assigning greater weight to ensemble models that exhibit the better performance. Three different numerical models, including European Center Medium-Range Weather Forecast, Ensemble Prediction System for Global, and Limited Area Ensemble Prediction System, were used to perform the post-processing method. We compared the model outputs from the weighed combination of ensembles with those from the Ensemble Model Output Statistics (EMOS) model for each raw ensemble model. The results showed that the weighted ensemble method can significantly improve the post-processing performance, compared to the raw ensemble method of the numerical models.

A Feature Selection-based Ensemble Method for Arrhythmia Classification

Erdenetuya Namsrai,Tsendsuren Munkhdalai,Meijing Li,Jung Hoon Shin,Oyun Erdene Namsrai,Keun Ho Ryu 한국정보처리학회 2013 Journal of information processing systems Vol.9 No.1

In this paper a novel method is proposed to build an ensemble of classifiers by using a feature selection schema. The feature selection schema identifies the best feature sets that affect the arrhythmia classification. Firstly a number of feature subsets are extracted by applying the feature selection schema to the original dataset. Then classification models are built by using the each feature subset. Finally we combine the classification models by adopting a voting approach to form a classification ensemble. The voting approach in our method involves both classification error rate and feature selection rate to calculate the score of the each classifier in the ensemble. In our method the feature selection rate depends on the extracting order of the feature subsets. In the experiment we applied our method to arrhythmia dataset and generated three top disjointed feature sets. We then built three classifiers based on the top-three feature subsets and formed the classifier ensemble by using the voting approach. Our method can improve the classification accuracy in high dimensional dataset. The performance of each classifier and the performance of their ensemble were higher than the performance of the classifier that was based on whole feature space of the dataset. The classification performance was improved and a more stable classification model could be constructed with the proposed approach.

Fault Detection of NPC Inverter Based on Ensemble Machine Learning Methods

Al-kaf Hasan Ali Gamal,Lee Jung-Won,Lee Kyo-Beum 대한전기학회 2024 Journal of Electrical Engineering & Technology Vol.19 No.1

Three-level neutral point clamped (NPC) inverters have been widely adopted in diferent appliances, but their growing use leads to increased susceptibility to faults in the system. It is therefore essential to design precise and efcient methods that can detect inverter faults to ensure optimal control and prevent serious damage to the system. However, the most accurate fault diagnosis methods often require signifcant amounts of time to collect input data such as current and voltage images, or they involve lengthy data rows that are not commonly applicable to real-time applications. To compensate for these drawbacks, ensemble machine learning (EML) methods are proposed to detect open-circuit faults that only require one single point as an input. Moreover, the proposed methods were trained using DC-link voltage diference, time, and three phase currents to improve the accuracy of open-circuit fault detection. The feasibility and efectiveness of the proposed method are verifed through simulation and experimentation. The present work also presents a comprehensive comparison of EML methods. The results show that Random Forest (RF) and Bootstrap Aggregating (bagging) methods achieve high performance compared to other EML methods, with an accuracy of 97%, without requiring additional circuitry. Additionally, the results show that incorporating time and DC-link voltage diferences, along with three-phase current, improves the performance of EML methods.

A Feature Selection-based Ensemble Method for Arrhythmia Classification

Namsrai, Erdenetuya,Munkhdalai, Tsendsuren,Li, Meijing,Shin, Jung-Hoon,Namsrai, Oyun-Erdene,Ryu, Keun Ho Korea Information Processing Society 2013 Journal of information processing systems Vol.9 No.1

In this paper, a novel method is proposed to build an ensemble of classifiers by using a feature selection schema. The feature selection schema identifies the best feature sets that affect the arrhythmia classification. Firstly, a number of feature subsets are extracted by applying the feature selection schema to the original dataset. Then classification models are built by using the each feature subset. Finally, we combine the classification models by adopting a voting approach to form a classification ensemble. The voting approach in our method involves both classification error rate and feature selection rate to calculate the score of the each classifier in the ensemble. In our method, the feature selection rate depends on the extracting order of the feature subsets. In the experiment, we applied our method to arrhythmia dataset and generated three top disjointed feature sets. We then built three classifiers based on the top-three feature subsets and formed the classifier ensemble by using the voting approach. Our method can improve the classification accuracy in high dimensional dataset. The performance of each classifier and the performance of their ensemble were higher than the performance of the classifier that was based on whole feature space of the dataset. The classification performance was improved and a more stable classification model could be constructed with the proposed approach.

Predicting Stock Liquidity by Using Ensemble Data Mining Methods

Bae, Eun Chan,Lee, Kun Chang 한국컴퓨터정보학회 2016 韓國컴퓨터情報學會論文誌 Vol.21 No.6

In finance literature, stock liquidity showing how stocks can be cashed out in the market has received rich attentions from both academicians and practitioners. The reasons are plenty. First, it is known that stock liquidity affects significantly asset pricing. Second, macroeconomic announcements influence liquidity in the stock market. Therefore, stock liquidity itself affects investors' decision and managers' decision as well. Though there exist a great deal of literature about stock liquidity in finance literature, it is quite clear that there are no studies attempting to investigate the stock liquidity issue as one of decision making problems. In finance literature, most of stock liquidity studies had dealt with limited views such as how much it influences stock price, which variables are associated with describing the stock liquidity significantly, etc. However, this paper posits that stock liquidity issue may become a serious decision-making problem, and then be handled by using data mining techniques to estimate its future extent with statistical validity. In this sense, we collected financial data set from a number of manufacturing companies listed in KRX (Korea Exchange) during the period of 2010 to 2013. The reason why we selected dataset from 2010 was to avoid the after-shocks of financial crisis that occurred in 2008. We used Fn-GuidPro system to gather total 5,700 financial data set. Stock liquidity measure was computed by the procedures proposed by Amihud (2002) which is known to show best metrics for showing relationship with daily return. We applied five data mining techniques (or classifiers) such as Bayesian network, support vector machine (SVM), decision tree, neural network, and ensemble method. Bayesian networks include GBN (General Bayesian Network), NBN (Naive BN), TAN (Tree Augmented NBN). Decision tree uses CART and C4.5. Regression result was used as a benchmarking performance. Ensemble method uses two types-integration of two classifiers, and three classifiers. Ensemble method is based on voting for the sake of integrating classifiers. Among the single classifiers, CART showed best performance with 48.2%, compared with 37.18% by regression. Among the ensemble methods, the result from integrating TAN, CART, and SVM was best with 49.25%. Through the additional analysis in individual industries, those relatively stabilized industries like electronic appliances, wholesale & retailing, woods, leather-bags-shoes showed better performance over 50%.

Uncertainty assessment of ensemble streamflow prediction method

Kim Seon-Ho,Kang Shin-Uk,Bae Deg-Hyo 한국수자원학회 2018 한국수자원학회논문집 Vol.51 No.6

본 연구에서는 충주댐 유역에 대해 앙상블 유량예측기법의 강우-유출 모델 매개변수, 입력자료에 따른 불확실성 분석을 수행하였다. 앙상블 유량예측기법으로는 ESP (Ensemble Streamflow Prediction) 기법과 BAYES-ESP (Bayesian-ESP) 기법을 활용하였으며, 강우-유출 모델로는 ABCD를 활용하였다. 모델 매개변수에 따른 불확실성 분석은 GLUE (Generalized Likelihood Uncertainty Estimation) 기법을 적용하였으며, 입력자료에 따른 불확실성 분석은 유량예측 앙상블에 활용되는 기상시나리오의 기간에 따라 수행하였다. 연구결과 앙상블 유량예측 기법은 입력자료 보다 모델 매개변수의 영향을 크게 받았으며, 20년 이상의 관측 기상자료가 확보되었을 때 활용하는 것이 적절하였다. 또한 BAYES-ESP는 ESP에 비해 불확실성을 감소시킬 수 있는 것으로 나타났다. 본 연구는 불확실성 분석을 통해 앙상블 유량예측기법의 특징을 규명하고 오차의 원인을 분석하였다는 점에서 가치가 있다고 판단된다. The objective of this study is to analyze uncertainties of ensemble-based streamflow prediction method for model parameters and input data. ESP (Ensemble Streamflow Prediction) and BAYES-ESP (Bayesian-ESP) based on ABCD rainfall-runoff model were selected as streamflow prediction method. GLUE (Generalized Likelihood Uncertainty Estimation) was applied for the analysis of parameter uncertainty. The analysis of input uncertainty was performed according to the duration of meteorological scenarios for ESP. The result showed that parameter uncertainty was much more significant than input uncertainty for the ensemble-based streamflow prediction. It also indicated that the duration of observed meteorological data was appropriate to using more than 20 years. And the BAYES-ESP was effective to reduce uncertainty of ESP method. It is concluded that this analysis is meaningful for elaborating characteristics of ESP method and error factors of ensemble-based streamflow prediction method.

거리기반 앙상블스무더를 이용한 채널저류층 불확실성평가

이경북,정승필,최종근 한국자원공학회 2015 한국자원공학회지 Vol.52 No.2

This paper suggests a new method of ensemble smoother(ES) for reliable uncertainty quantification. The method uses several Kalman gains rather than one representative Kalman gain. When the proposed method is applied to channelized reservoirs, the results manage typical overshooting and filter divergence problems. Also, they conserve channel connectivity and bimodal distribution of the model parameter. The proposed method can keep history matching time short since there are no modifications in the standard ES. Therefore, the time of the proposed method reduces more than 97% of that of ensemble Kalman filter(EnKF) with 45 assimilation steps and 200 total ensembles. The ES with a distance-based method provides reliable productions with reasonable uncertainty ranges. Also, prediction time of future performances can be reduced since the representative ensembles from each group estimate similar uncertainty ranges over all ensembles. Therefore, the proposed method can be applied for decision making because it gives fast and reliable uncertainty quantification for channelized reservoirs. 본 연구에서는 앙상블스무더의 불확실성평가 신뢰도를 향상하기 위해 초기앙상블을 잘 대표하는 다수의 칼만게인을 이용하는 기법을 제안하였다. 제안된 거리기반 앙상블스무더를 채널저류층에 적용한 결과, 기존의 오버슈팅과 필터발산 문제를 해결하였고 채널연결성과 이봉분포의 특징을 잘 보존하였다. 제안된 기법은앙상블스무더의 수식과 교정방식을 수정없이 사용하므로 계산속도가 빠르다. 따라서 200개의 앙상블로 45번의교정을 수행한 경우, 앙상블칼만필터의 소요시간보다 97% 이상 감소시켰다. 거리기반 앙상블스무더의 경우 유정별 생산량뿐만 아니라 누적 오일 및 물 생산량을 성공적으로 예측하였고 편향없는 불확실성을 제공하였다. 또한 군집별 대표앙상블만으로 전체앙상블과 비슷한 수준의 불확실성평가가 가능하므로 군집수에 반비례하여소요시간이 줄어든다. 따라서 제안된 기법은 빠르고 신뢰할 수 있는 불확실성평가가 가능하므로 채널저류층개발 시 의사결정을 위한 도구로 활용될 수 있다.

배깅 및 스태킹 기반 앙상블 기계학습법을 이용한 고성능 콘크리트 압축강도 예측모델 개발

곽윤지,고채연,곽신영,임승현 한국전산구조공학회 2023 한국전산구조공학회논문집 Vol.36 No.1

Predicting the compressive strength of high-performance concrete (HPC) is challenging because of the use of additional cementitious materials; thus, the development of improved predictive models is essential. The purpose of this study was to develop an HPC compressive-strength prediction model using an ensemble machine-learning method of combined bagging and stacking techniques. The result is a new ensemble technique that integrates the existing ensemble methods of bagging and stacking to solve the problems of a single machine-learning model and improve the prediction performance of the model. The nonlinear regression, support vector machine, artificial neural network, and Gaussian process regression approaches were used as single machine-learning methods and bagging and stacking techniques as ensemble machine-learning methods. As a result, the model of the proposed method showed improved accuracy results compared with single machine-learning models, an individual bagging technique model, and a stacking technique model. This was confirmed through a comparison of four representative performance indicators, verifying the effectiveness of the method. 고성능 콘크리트(HPC) 압축강도는 추가적인 시멘트질 재료의 사용으로 인해 예측하기 어렵고, 개선된 예측 모델의 개발이 필수적이다. 따라서, 본 연구의 목적은 배깅과 스태킹을 결합한 앙상블 기법을 사용하여 HPC 압축강도 예측 모델을 개발하는 것이다. 이 논문의 핵심적 기여는 기존 앙상블 기법인 배깅과 스태킹을 통합하여 새로운 앙상블 기법을 제시하고, 단일 기계학습 모델의 문제점을 해결하여 모델 예측 성능을 높이고자 한다. 단일 기계학습법으로 비선형 회귀분석, 서포트 벡터 머신, 인공신경망, 가우시안 프로세스 회귀를 사용하고, 앙상블 기법으로 배깅, 스태킹을 이용하였다. 결과적으로 본 연구에서 제안된 모델이 단일 기계학습 모델, 배깅 및 스태킹 모델보다 높은 정확도를 보였다. 이는 대표적인 4가지 성능 지표 비교를 통해 확인하였고, 제안된 방법의 유효성을 검증하였다.