시계열 교차검증을 적용한 2,3-BDO 분리공정 온도예측 모델의 초매개변수 최적화

안나현 ( Nahyeon An ),최영렬 ( Yeongryeol Choi ),조형태 ( Hyungtae Cho ),김정환 ( Junghwan Kim ) 한국화학공학회 2021 Korean Chemical Engineering Research(HWAHAK KONGHA Vol.59 No.4

최근 인공지능에 대한 관심이 높아짐에 따라 화학공정분야에서도 인공지능을 활용한 연구가 많아지고 있다. 그러나 인공지능 기반 모델이 충분히 일반화되지 않아 학습에 이용되지 않은 새로운 데이터에 대한 예측률이 떨어지는 과적합 현상이 빈번하게 일어나고 있으며, 교차검증은 과적합을 해결하는 방법 중 하나이다. 본 연구에서는 2,3-BDO 분리 공정 온도 예측 모델의 초매개변수 중에서 배치 개수와 반복횟수를 조정하기 위해 시계열 교차검증을 적용하고 일반적으로 사용되는 K 겹 교차검증과 비교하였다. 결과적으로 K 겹 교차검증을 사용했을 때 보다 시계열 교차검증 방식을 사용했을 때 MAPE는 0.61% 증가한 반면 RMSE는 9.06% 감소하였고 학습 시간은 198.29초 적게 소요되었다. Recently, research on the application of artificial intelligence in the chemical process has been increasing rapidly. However, overfitting is a significant problem that prevents the model from being generalized well to predict unseen data on test data, as well as observed training data. Cross validation is one of the ways to solve the overfitting problem. In this study, the time-series cross validation method was applied to optimize the number of batch and epoch in the hyperparameters of the prediction model for the 2,3-BDO distillation process, and it compared with K-fold cross validation generally used. As a result, the RMSE of the model with time-series cross validation was lower by 9.06%, and the MAPE was higher by 0.61% than the model with K-fold cross validation. Also, the calculation time was 198.29 sec less than the K-fold cross validation method.

친환경공급사슬관리 측정모델의 개발과 교차타당성 검증

주혜영(Hye-Young Joo) 한국물류학회 2014 물류학회지 Vol.24 No.5

본 연구는 한국과 중국에서 동일하게 기능할 수 있는 Green supply chain management 측정모델을 개발하는 것을 주요한 목적으로 한다. 이를 위해 대표적 실행모델이라고 할 수 있는 Vachon and Klassen(2006)b, Zhu and Sarkis(2006)의 연구모델과 본 제안모델을 교차타당성과 간명성을 중심으로 평가하였다. 연구의 결과는 다음과 같다. 첫째, Vachon and Klassen(2006)b, Zhu and Sarkis(2006) 및 본 연구모델은 모두 교차타당성이 성립하는 것으로 나타났다. 둘째, Vachon and Klassen(2006)b 의 연구모델은 교차타당성의 측면에서는 가장 우수하나 적합도는 상대적으로 낮았으며 Zhu and Sarkis(2006)의 모델은 적합도는 우수했으나 교차타당성은 다소 낮게 나타났다. 본 제안모델의 경우는 모델의 적합도가 높았고 교차타당성도 상대적으로 우수한 것으로 나타났다. 따라서 본 모델은 Green SCM 측정모델의 교차타당성과 간명성을 높인 모델로 평가할 수 있다. The purpose of this study is to develop a measurement model of green supply chain management, which can perform the same functions between Korea and China. To do this, we evaluated Vachon & Klassen(2006)b, Zhu & Sarkis(2006), and the proposed model proposed in respect of cross validation and parsimony. The results can be summarized as follows: First, all research models of Vachon and Klassen(2006)b, Zhu and Sarkis(2006) and the proposed model accomplished a cross validation within a research model. Second, Vachon and Klassen(2006)b research model was superior to the others in respect of cross validation, but not that of goodness of fit. Zhu and Sarkis(2006)’s research model was excellent in goodness of fit of model, while cross validation was not so. In the case of the proposed model, the model’s excellence in the goodness of fit and cross validation were all higher than the others’ level. Thus, this research showed that the proposed model’s excellence in respect of cross validation and parsimony.

SVM 교차검증을 활용한 토지피복 ROI 선정

정종철,윤형진 한국국토정보공사 2020 지적과 국토정보 Vol.50 No.1

This study examines machine learning cross-validation to utilized create ROI for classification of land cover. The study area located in Sejong and one KOMPSAT-3A image was used in this analysis: procedure on October 28, 2019. We used four bands(Red, Green, Blue, Near infra-red) for learning cross validation process. In this study, we used K-fold method in cross validation and used SVM kernel type with cross validation result. In addition, we used 4 kernels of SVM(Linear, Polynomial, RBF, Sigmoid) for supervised classification land cover map using extracted ROI. During the cross validation process, 1,813 data extracted from 3,500 data, and the most of the building, road and grass class data were removed about 60% during cross validation process. Based on this, the supervised SVM linear technique showed the highest classification accuracy of 91.77% compared to other kernel methods. The grass’ producer accuracy showed 79.43% and identified a large mis-classification in forests. Depending on the results of the study, extraction ROI using cross validation may be effective in forest, water and agriculture areas, but it is deemed necessary to improve the distinction of built-up, grass and bare-soil area. 본 연구는 토지피복 분류에 사용 가능한 ROI 생성 과정에서 기계학습 기반 교차검증을 활용하였다. 연구지역은 세종시를 포함한 2019년 10월 28일 단시기 KOMPSAT-3A 영상을 활용하였다. 연구 과정에서 4개의 밴드(Red, Green, Blue, Near Infra-red)를 독립변수로 교차검증 과정에서 학습시켰다. 또한 SVM의 4가지 기법(Linear, Polynomial, RBF, Sigmoid)을 활용하여 추출된 ROI를 기반으로 토지피복 분류를 실시하였다. 교차검증 과정에서 훈련된 3,500개의 데이터 중 1,813개의 데이터가 추출되었으며 건물, 도로, 그리고 초지에서 약 60%의 데이터가 제거되었다. 추출된 ROI를 기반으로 다른 SVM기법에 비해 SVM Linear 기법이 91.77%로 가장 높은 분류 정확도를 나타냈다. 분류 클래스 중 초지의 경우 산림과의 오분류가 가장 많이 발생하며 79.43%의 생산자 정확도로 가장 낮은 분류 정확도를 보여주었다. 연구 결과에 따라 교차검증에서 추출된 ROI는 산림, 수역, 그리고 농업지역에 대해서는 90%이상의 분류정확도를 보여주며 효과적인 분류결과를 도출할 수 있었으나, 80%의 분류정확도를 보여주는 건물, 도로, 나대지, 그리고 초지 지역을 분류하는 방법에 대해서는 추가적인 연구가 진행되어야 할 필요성이 존재한다.

소변 중 다환방향족탄화수소 대사체의 분석법 확립 및 교차분석

박나연,전중대,구혜령,김정환,이은희,이경무,문철진,고영림,Park, Na-Youn,Jeon, Jung-Dae,Koo, Hyeryeong,Kim, Jung Hoan,Lee, Eun-Hee,Lee, Kyungmu,Mun, Cheoljin,Kho, Younglim 한국환경보건학회 2015 한국환경보건학회지 Vol.41 No.5

Objectives: This study was performed to evaluate the analytical method for PAH metabolites in human urine using enzyme hydrolysis and solid-phase extraction coupled with LC-(ESI)-MS/MS technique. Methods: We employed HPLC tandem mass spectrometry techniques with appropriate pre-treatment for analysis of 16 OH-PAHs in human urine. Samples were hydrolysis by ${\beta}$-flucuronidase/Aryl sulfatase, and target compounds were extracted by solid-phase extraction with a strata-x cartridge. Cross-validation was performed between Eulji University and Green Cross laboratories with 200 human urine samples. Results: The accuracies were between 90.3% and 118.8%, and precisions (relative standard deviations) were lower than 10%. The linearity obtained was satisfying for the 16 OH-PAH compounds, with a coefficient of determination ($r^2$) higher than 0.99. The results of cross-validation at the two organizations were compared by ICC (interclass correlation coefficient) values. The cross-validation results were excellent or good for all compounds. Conclusion: An analytical method was validated for low nanogram levels of 16 OH-PAHs in human urine. Also, satisfying results were obtained for method validation such as accuracy, precision and ICC of cross-validation.

Robust Cross Validations in Ridge Regression

Kang-Mo Jung 한국전산응용수학회 2009 Journal of applied mathematics & informatics Vol.27 No.3

The shrink parameter in ridge regression may be contaminated by outlying points. We propose robust cross validation scores in ridge regression instead of classical cross validation. We use robust location estimators such as median, least trimmed squares, absolute mean for robust cross validation scores. The robust scores have global robustness. Simulations are performed to show the effectiveness of the proposed estimators. The shrink parameter in ridge regression may be contaminated by outlying points. We propose robust cross validation scores in ridge regression instead of classical cross validation. We use robust location estimators such as median, least trimmed squares, absolute mean for robust cross validation scores. The robust scores have global robustness. Simulations are performed to show the effectiveness of the proposed estimators.

Development and Cross-Validation of Equation for Estimating Percent Body Fat of Korean Adults According to Body Mass Index

성호용,문준배 대한비만학회 2017 The Korean journal of obesity Vol.26 No.2

Background: Using BMI as an independent variable is the easiest way to estimate percent body fat. Thus far, few studies have investigated the development and cross-validation of an equation for estimating the percent body fat of Korean adults according to the BMI. The goals of this study were the development and cross-validation of an equation for estimating the percent fat of representative Korean adults using the BMI. Methods: Samples were obtained from the Korea National Health and Nutrition Examination Survey between 2008 and 2011. The samples from 2008-2009 and 2010-2011 were labeled as the validation group (n=10,624) and the cross-validation group (n=8,291), respectively. The percent fat was measured using dual-energy X-ray absorptiometry, and the body mass index, gender, and age were included as independent variables to estimate the measured percent fat. The coefficient of determination (R2), standard error of estimation (SEE), and total error (TE) were calculated to examine the accuracy of the developed equation. Results: The cross-validated R2 was 0.731 for Model 1 and 0.735 for Model 2. The SEE was 3.978 for Model 1 and 3.951 for Model 2. The equations developed in this study are more accurate for estimating percent fat of the cross-validation group than those previously published by other researchers. Conclusion: The newly developed equations are comparatively accurate for the estimation of the percent fat of Korean adults.

A Study of Model Selection for Electric Data using Cross Validation Approach

Saraswathi Sivamani,Saravana Kumar,신창선,박장우,조용윤 한국지식정보기술학회 2017 한국지식정보기술학회 논문지 Vol.12 No.6

In this paper, the appropriate model is selected for the risk assessment of the electric utility pole data with the help of cheat sheets and k-fold cross validation. In order to analyze, predict and forecast the data, the appropriate model has to be selected. The major issue is the declination of the accuracy in the model fitting, which may result in poor model selection. There are different type of machine learning algorithm, which makes it difficult to conclude the model selection. To ensure the proper selection of the model, we undergo a two-step process. Firstly, the basic model is selected with the existing model selection cheat sheets named as Scikit learn and Microsoft azure, by understanding the available input and required output of the data. After getting through the multiple question, the respective models such as Generalized Additive Model, Generalized Linear Model, Linear Regression and Support Vector Machine are obtained. In order to attain the appropriate model, we perform k-fold cross validation to estimate the risk of the algorithms, by comparing 2-fold, 8-fold and 10-fold cross validation. Between the three set, the 10-cross fold validation of generalized additive model is selected with the least risk error. Using k-fold cross validation, we estimate the accuracy of the model that is suitable for the data, by using the electric power data set.

K겹 교차검증 및 심층신경망을 활용한 자동차 시트의 BSR 소음 지표 예측 연구

김석범,남재현,고동신 한국자동차공학회 2024 한국 자동차공학회논문집 Vol.32 No.1

This study proposes a method for predicting Loudness N10, a quantitative indicator for evaluating BSR noise in automotive seats. The approach utilizes k-fold cross-validation and deep neural networks(DNNs) to predict the indicator without expensive equipment or specific software. Experimental data on acoustic and sound quality physical quantities were obtained, with significant factors such as sound pressure level and variation intensity identified. While linear and nonlinear regression equations using k-fold cross-validation resulted in large prediction errors, the DNN-based prediction model demonstrated lower errors. The integration of k-fold cross-validation helps maintain performance in limited environments. In summary, the proposed method enables accurate prediction of Loudness N10 based on acoustic and sound quality parameters, even in resource-constrained settings.

순차적 크리깅모델의 평균-분산 정확도 검증기법

이태희(Tae Hee Lee),김호성(Hosung Kim) 대한기계학회 2010 大韓機械學會論文集A Vol.34 No.5

메타모델의 정확도를 엄밀하게 검증하는 것은 메타모델링에서 중요한 연구주제이다. k 점 선택교차검증기법이 많은 계산시간을 요구하면서도 메타모델의 정확도를 정략적으로 측정하지 못한다. 최근들어, 평균 ? 기준이 메타모델의 정확도를 정량적으로 제공하기 위하여 제안되었다. 그러나 평균 ? 검증 기준은 크리깅 메타모델이 부정확함에도 불구하고 일찍 수렴하는 경향이 있다. 따라서 본 연구에서는 최대엔트로피를 이용한 순차적 실험계획에서 크리깅모델의 평균과 분산을 이용한 정확도 평가기법을 제안한다. 이 제안한 기법은 평균 및 분산을 계산할 때 수치해석으로 구하는 것이 아니라 크리깅메타모델을 직접 적분하여 구하기 때문에 k 점 선택교차검증기법보다 효율적이며 정확하다. 제안한 기준은 실제 응답의 평균제곱오차의 경향과 매우 유사하여 순차적 실험계획의 수렴기준으로 사용할 수 있다. The rigorous validation of the accuracy of metamodels is an important topic in research on metamodel techniques. Although a leave-k-out cross-validation technique involves a considerably high computational cost, it cannot be used to measure the fidelity of metamodels. Recently, the mean? validation technique has been proposed to quantitatively determine the accuracy of metamodels. However, the use of mean? validation criterion may lead to premature termination of a sampling process even if the kriging model is inaccurate. In this study, we propose a new validation technique based on the mean and variance of the response evaluated when sequential sampling method, such as maximum entropy sampling, is used. The proposed validation technique is more efficient and accurate than the leave-k-out cross-validation technique, because instead of performing numerical integration, the kriging model is explicitly integrated to accurately evaluate the mean and variance of the response evaluated. The error in the proposed validation technique resembles a root mean squared error, thus it can be used to determine a stop criterion for sequential sampling of metamodels.

잠재평균분석을 활용한 태권도 선수의 2×2성취목표성향 차이 분석

염대관 ( Daekwan Yeom ),김창호 ( Changho Kim ),김우진 ( Woojin Kim ) 대한무도학회 2015 대한무도학회지 Vol.17 No.3

본 연구는 중·고등학교 및 대학교 태권도 겨루기 선수를 중심으로 성별에 따른 2×2성취목표성향 차이를 분석하는데 그 목적이 있다. 이를 위해 중고교, 대학교 태권도 선수 187명을 표집 하였으며 잠재평균 분석을 실시하였다. 잠재평균분석을 적용하기 전에 수렴타당도와 판별타당도 그리고 교차 타당성 검증을 실시하였다. 수렴타당도 및 판별타당도는 탐색적 요인분석과 확인적 요인분석을 적용하였으며 교차타당성 검증을 위해 다중집단 확인적 요인분석을 실시하였다. 구체적으로 그 내용을 살펴보면 다음과 같다. 탐색적 요인분석을 통해 구성타당도를 살펴본 결과 선행연구와 동일하게 4요인 모형이 채택되었으며 적합도(RMSEA)가 그 기준을 만족하였다. 수렴타당도 및 판별타당도를 확인하기 위하여 확인적 요인분석을 실시하였다. 또한 표준화 계수, 평균분산추출값, 개념신뢰도를 확인한 결과 그 기준을 만족하였으며 판별타당도 검증 결과에서도 평균분산추출값이 상관계수의 제곱값보다 큰 것으로 나타나 모든 기준을 만족하였다. 잠재평균분석 전 교차타당성 검증을 위해 형태동일성, 측정동일성, 절편동일성 및 요인분산동일성을 검증하였다. 이를 위해 각 모형의 적합도를 확인하였으며 모형간 적합도 차이를 분석하였다. 마지막으로 잠재평균분석 결과 수행회피는 여성이 남성에 비해 높은 것으로 나타났으며 수행접근은 여성에 비해 남성이 높은 것으로 숙달회피는 여성에 비해 남성이 높은 것으로 나타났으며 통계적으로 유의하게 분석되었다. 숙달접근 역시 여성에 비해남성이 더 높은 것으로 나타났으나 통계적으로 유의하지 않았다. The purpose of this study is to analyze the difference of 2x2 achievement goal orientation along the sex focusing on Taekwondo competition athletes of middle school, high school and university. 187 middle school, high school and university Taekwondo athletes were collected as the sample for that and potential average analysis was carried out. convergence validity, discriminant validity and cross-validation were verified before applying potential average analysis. The exploratory factor analysis and confirmatory factor analysis were applied to convergence validity and discriminant validity and confirmatory factor analysis of multiple group was carried out to verify cross-validation. Looking into the contents in detail, the results are as following. As the results of looking into construct validity through exploratory factor analysis, 4-factors model was selected as the same as the advanced research and Root Mean Square Error of Approximation(RMSEA) satisfied the standard. The confirmatory factor analysis was carried out to check convergence validity and discriminant validity. And as the results of checking standardization coefficient, average variance extract value and construct reliability, the standard was satisfied and it satisfied all standards as it appeared that average variance extract was larger than the square value of coefficient of correlation the results of verifying discriminant validity as well. In order to verify cross-validation before potential average analysis, configural invariance, measurement invariance, specimen invariance and factor variance invariance were verified. The fitness of each model was checked for that and the difference of fitness among the models was analyzed. Lastly, as the result of potential average analysis, it appeared that female showed the higher value than male regarding performance-avoidance, male showed the higher value than female regarding performance-access, female showed the higher value regarding proficiency-avoidance and it was analyzed as significantly meaningful. Male showed the higher value than female regarding proficiency-avoidance as well, however, it was not significantly meaningful.