• 한국경영과학회지
• /
• 제20권3호
• /
• pp.17-29
• /
• 1995

Recently, neural network models have been employed as an alternative to regression analysis for point estimation or function fitting in various field. Thus far, however, no theoretical or empirical guides seem to exist for selecting the tool which the most suitable one for a specific function-fitting problem. In this paper, we evaluate performance of three major function-fitting techniques, regression analysis and two neural network models, back-propagation and linear-Hebbian-learning neural networks. The functions to be fitted are simple linear ones of a single independent variable. The factors considered are size of noise both in dependent and independent variables, portion of outliers, and size of the data. Based on comutational results performed in this study, some guidelines are suggested to choose the best technique that can be used for a specific problem concerned.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 1987년도 제29회 수공학연구발표회논문초록집
• /
• pp.151-158
• /
• 1987

The purpose of thes study is to estimate the parameters of linear reservoir models in order to derive the instantaneous unit hydrograph from a given small experimental watershed. The linear reservoir model is a conceptual model, consisting of cascade or parallel equal linear reservoirs, preceded by a linear channel which involved NASH, SLR(single linear reservoir)and 2-PLR(two-parallel linear reservoir)model. The NASH model have two parameters N and K, single linear reservoir has one parameter K1 and two-parallel linear reservoirs have two parameters K1, K2;where N denote the number of reservoirs and K is the storage coefficient of each reservoirs.

• 전자공학회논문지B
• /
• 제33B권11호
• /
• pp.22-30
• /
• 1996

In this paper, the properties and the design procedures of the internal model control (IMC) structures are discussed and a new nonlinear IMC(NIMC) strategy is proposed. The IMC controllers are simply inverse controller in principle but the development of a NIMC poses difficulties due to the inherent complexity of nonlinear systems. Existing design mehtods are a few and not easy to implement. The proposed approach is using multiple linear models, linear IMC controllers, and swiching scheme instead of using nonlinear model/controller. The advantages of the new approach are that we can use linear IMC mehtod which are now well estabilished and need not global nonlinear models.

• 전기학회논문지
• /
• 제60권8호
• /
• pp.1622-1626
• /
• 2011

Recently, interests on microgrids have been growing as clean power systems. Microgrids include small scaled distributed generation such as wind and solar power as well as diesel generators as main power sources. To operate a microgrid effectively, optimal scheduling for the microgrid is important. Especially, in the grid-connected mode, power trades between the microgrid and the power grid should be considered in optimal scheduling. In this paper, mathematic models for optimal operation of a microgrid were established based on the linear programming. In particular, the shiftable load was considered in the models to optimize it in microgrid operation. To show feasibility of the proposed models, they were applied to optimal microgrid operation and the results were discussed.

• Communications for Statistical Applications and Methods
• /
• 제21권5호
• /
• pp.423-433
• /
• 2014

Test statistics using cumulative sums of residuals have been widely used in various regression models including generalized linear models(GLM). Recently, Pan and Lin (2005) extended this testing procedure to the generalized linear mixed models(GLMM) having random effects, in which we encounter difficulties in computing the marginal likelihood that is expressed as an integral of random effects distribution. The Gaussian quadrature algorithm is commonly used to approximate the marginal likelihood. Many commercial statistical packages provide an option to apply this type of goodness-of-fit test in GLMs but available programs are very rare for GLMMs. We suggest a computational algorithm to implement the testing procedure in GLMMs by a freely accessible R package, and also illustrate through practical examples.

• Journal of Electrical Engineering and Technology
• /
• 제11권2호
• /
• pp.265-273
• /
• 2016

A water wall system is one of the most important components of a boiler in a thermal power plant, and it is a nonlinear Multi-Input and Multi-Output (MIMO) system, with 6 inputs and 3 outputs. Three models are developed and comp for the controller design, including a linear model, a multilayer feed-forward neural network (MFNN) model and an Echo State Network (ESN) model. First, the linear model is developed by linearizing a given nonlinear model and is analyzed as a function of the operating point. Second, the MFNN and the ESN are developed by using training data from the nonlinear model. The three models are validated using Matlab with nonlinear input-output data that was not used during training.

• 응용통계연구
• /
• 제18권2호
• /
• pp.395-410
• /
• 2005

많은 수의 범주형 변수에 의한 대규모 분할표 분석을 위하여 차원축소(collapsibility) 성질을 이용한 분석 방법을 제안하였다. kullback-Leibler의 발산 측도(divergence measure)를 이용한 서로 완전한 연관을 갖는 변수그룹을 결정하는 방법을 제안하였으며, 제안된 방법에 의한 변수그룹은 주변 로그선형모형(marginal log-linear models)에 의하여 변수들간의 연관성을 식별할 수 있다. 제안된 방법의 적용 예로 데이터 마이닝에서 흔히 접할 수 있는 대규모 분할표 자료인 소비자들의 구매행위 분석을 위한 장바구니 자료의 분석 결과를 제시하였다.

• 대한용접접합학회:학술대회논문집
• /
• 대한용접접합학회 2006년 추계학술발표대회 개요집
• /
• pp.271-273
• /
• 2006

This paper focuses on the developed empirical models for the prediction on top-bead width in GMA(Gas Metal Arc) welding process. Three empirical models have been developed: linear, curvilinear and an intelligent model. Regression analysis was employed fur optimization of the coefficients of linear and curvilinear model, while Genetic Algorithm(GA) was utilized to estimate the coefficients of intelligent model. Not only the fitting of these models were checked, but also the prediction on top-bead width was carried out. ANOVA analysis and contour plots were respectively employed to represent main and interaction effects between process parameters on top-bead width.

• 대한화학회지
• /
• 제58권6호
• /
• pp.543-552
• /
• 2014

A quantitative structure activity relationship (QSAR) study is performed for modeling and prediction of oral bioavailabilities of 216 diverse set of drugs. After calculation and screening of molecular descriptors, linear and nonlinear models were developed by using multiple linear regression (MLR), artificial neural network (ANN), support vector machine (SVM) and random forest (RF) techniques. Comparison between statistical parameters of these models indicates the suitability of SVM over other models. The root mean square errors of SVM model were 5.933 and 4.934 for training and test sets, respectively. Robustness and reliability of the developed SVM model was evaluated by performing of leave many out cross validation test, which produces the statistic of $Q^2_{SVM}=0.603$ and SPRESS = 7.902. Moreover, the chemical applicability domains of model were determined via leverage approach. The results of this study revealed the applicability of QSAR approach by using SVM in prediction of oral bioavailability of drugs.

• 품질경영학회지
• /
• 제23권3호
• /
• pp.113-125
• /
• 1995

This paper describes the 2-stage randomized response model in the Bayesian view point. The classical Bayesian analysis needs the complete information for a prior density, but the Bayes linear estimator needs only the first and second moments. Therefore, it is convenient to find the estimator and this estimator robusts to a prior density. We show that MSE's of the Bayes linear estimators for the 2-stage randomized response models are smaller than those of the MLE's for the 2-stage randomized response models.