• Journal of the Korean Institute of Gas
• /
• v.16 no.1
• /
• pp.26-32
• /
• 2012

The autoignition temperature is one of the most important physical properties used to determine the flammability characteristics of chemical substances. Despite the needs of the experimental autoignition temperature data for the design of chemical plants, it is not easy to get the data. This study have built and compared partial least squares (PLS) and support vector machine (SVM) models to predict the autoignition temperatures of 503 organic compounds out of DIPPR 801. As the independent variables of the models, 59 functional groups were chosen based on the group contribution method. The prediction errors calculated from cross-validation were employed to determine the optimal parameters of two models. And, particle swarm optimization was used to get three parameters of SVM model. The PLS and SVM results of the average absolute errors for the whole data range from 58.59K and 29.11K, respectively, indicating that the predictive ability of the SVM is much superior than PLS.

• Clean Technology
• /
• v.18 no.1
• /
• pp.31-37
• /
• 2012

Biomass has become a major feedstock for bioenergy and other bio-based products because of its renewability and environmental benefits. Various researches have been done in the prediction of crucial characteristics of biomass, including the active utilization of spectroscopy data. Near infrared (NIR) spectroscopy has been widely used because of its attractive features: it's non-destructive and cost-effective producing fast and reliable analysis results. This work developed the multivariate statistical scheme for predicting weight loss profiles based on the utilization of NIR spectra data measured for six lignocellulosic biomass types. Wavelet analysis was used as a compression tool to suppress irrelevant noise and to select features or wavelengths that better explain NIR data. The developed scheme was demonstrated using real NIR data sets, in which different prediction models were evaluated in terms of prediction performance. In addition, the benefits of using right pretreatment of NIR spectra were also given. In our case, it turned out that compression of high-dimensional NIR spectra by wavelet and then PLS modeling yielded more reliable prediction results without handling full set of noisy data. This work showed that the developed scheme can be easily applied for rapid analysis of biomass.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2000.05a
• /
• pp.406-409
• /
• 2000

We propose the modified neural network which are paralleled to control nonlinear systems. The proposed method is a direct control method to use inverse model of the plant. Nonlinear systems are divided into two parts; linear part and nonlinear part, and it is controlled by RLS method and recursive multi-layer neural network with each other. We simulate to verify the performance of the proposed method and are compared with conventional direct neural network control method. The proposed control method is improved the control performance than the conventional method.

• Proceedings of the KIEE Conference
• /
• 2015.07a
• /
• pp.1375-1376
• /
• 2015

본 논문은 퍼지 클러스터링을 이용한 규칙 성장 기반 퍼지 분류기의 설계에 대해서 소개한다. 본 논문의 목적은 퍼지 클러스터링을 통해 형성된 증가된 퍼지 규칙을 이용한 새로운 설계 방법론을 개발하는 것이다. 제안된 분류기는 네개의 기능적인 부분으로 구성된다. 퍼지 규칙의 전반부는 퍼지 클러스터링 알고리즘을 이용해 구성된 멤버쉽 함수를 나타낸다. 후반부는 지역 모델을 구성한다. 지역 모델의 파라미터는 가중 최소 자승법에 의해 추정된다. 추론부에서는, 각 퍼지 규칙의 에러 측정후, 가장 높은 에러를 갖는 하나의 퍼지 규칙이 선택된다. 규칙성장 부분에서는, 네트워크의 강화를 위해 규칙의 성장 과정이 이루어지며, 선택된 규칙은 제안된 분류기에서 더 나은 성능을 위해 두 개 또는 세 개의 세분화된 퍼지 규칙으로 나누어진다. 이러한 새로운 규칙은 context 기반 Fuzzy C-Means 클러스터링에 의해서 형성된다. 제안된 규칙 기반 분류기의 효용성을 토론하며, 머신 러닝 데이터를 이용하여 실험을 수행하였다.

• Journal of the Society of Naval Architects of Korea
• /
• v.35 no.2
• /
• pp.20-28
• /
• 1998

To accurately predict the motion of a submergible, the nonlinear structure of dynamic model should be selected and corresponding parameters should be estimated using model test. Providing the model structure, only the values of parameters are unknown and the estimation can thus be formulated as a standard least square problem. Unfortunately, the nonlinear model structure of submersibles is rarely known a prior and method of model structure determination from measurement data of model test should be developed and included as a vital part of the estimation procedure. In this study, the well-known linear least square algorithm for the analysis of model tests and a way to measure the goodness are reviewed, and the identification algorithm based on an orthogonal decomposition method of Gram-Schmidt is extended to combine structure selection and maneuvering coefficients estimation in a very simple and efficient manner. Finally, the efficiency of this algorithm is verified by using simulation and applying to the analysis of model test of a submerged body. As a result, it was verified that this combined method might be very erective in selecting the structure of dynamic model estimating the maneuvering coefficients from measurement fiat of model test.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.35 no.1
• /
• pp.1-11
• /
• 2015

In this study, an online seed potato sorting system using a visible and near infrared (40 1100 nm) transmittance spectral technique and statistical model was evaluated for the nondestructive determination of infected and sound seed potatoes. Seed potatoes that had been artificially infected with Pectobacterium atrosepticum, which is known to cause a soil borne disease infection, were prepared for the experiments. After acquiring transmittance spectra from sound and infected seed potatoes, a determination algorithm for detecting infected seed potatoes was developed using the partial least square discriminant analysis method. The coefficient of determination($R^2_p$) of the prediction model was 0.943, and the classification accuracy was above 99% (n = 80) for discriminating diseased seed potatoes from sound ones. This online sorting system has good potential for developing a technique to detect agricultural products that are infected and contaminated by pathogens.

• Journal of Environmental Science International
• /
• v.6 no.5
• /
• pp.541-550
• /
• 1997

A least-square regression analysis is applied for the estimation of velocity streamfunction field based on discretely sampled current meter data. The coefficients of a streamfuunction that is expanded in terms of trigonometric basis function are obtained by enforcing the horizontal non-divergence of two-dimensional flow field. This method avoids Interpolation and gives a root-mean-square (rms) residual of fit which Includes the divergent part and noisiness of oceanic data. The Implementation of the method Is done by employing a boundary-fitted, curvilinear orthogonal coordinate which facilitates the specification of boundary conditions. An application is successfully made to the Texas-Louisiana shelf using the 32 months current meter data (31 moorings) observed as a part of the Texas-Louisiana Shelf and Transport Processes Study (LATEX). The rms residual of the fitting is relatively small for the shelf, which indicates the field Is Ivell represented by the streamnfunction.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.31 no.3
• /
• pp.31-38
• /
• 2003

In the present work, a nonlinear inverse perturbation method which has been used in the structural optimization, is adopted so as to identify the structural damages. Unlike the structural optimization, a larger number of constrained equations than the number of unknown parameters are often required detect structural damage. Therefore, nonlinear least squares method is utilized to solve the problem. Because only a limited number of sensors are available I real situation of damage detection, the determination of sensor location becomes one of the most important issues. Hence, this work concentrates on the issue of sensor placement in the framework of nonlinear inverse perturbation method, and the performances of various methodologies concerning to sensor placement are compared with each other. The comparisons show tat the successive elimination method gets good performance for sensor placement. From the several numerical studies, it is confirmed that the inverse perturbation method, combined with the successive elimination method, is very promising in structural damage detection.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.40 no.4
• /
• pp.264-274
• /
• 2003

This paper presents a direct generalized minimum-variance self tuning controller with a PID structure using neural network which adapts to the changing parameters of the nonlinear system with nonminimum phase behavior, noises and time delays. The self-tuning controller with a PID structure is a combination of the simple structure of a PID controller and the characteristics of a self-tuning controller that can adapt to changes in the environment. The self-tuning control effect is achieved through the RLS (recursive least square) algorithm at the parameter estimation stage as well as through the Robbins-Monro algorithm at the stage of optimizing the design parameter of the controller. The neural network control effect which compensates for nonlinear factor is obtained from the learning algorithm which the learning error between the filtered reference and the auxiliary output of plant becomes zero. Computer simulation has shown that the proposed method works effectively on the nonlinear nonminimum phase system with time delays and changed system parameter.

• Journal of Institute of Control, Robotics and Systems
• /
• v.10 no.2
• /
• pp.185-191
• /
• 2004

Partial least squares(PLS) is one of multiplicate statistical process methods and has been developed in various algorithms with the characteristics of principal component analysis, dimensionality reduction, and analysis of the relationship between input variables and output variables. But it has been limited somewhat by their dependency on linear mathematics. The algorithm is proposed to classify for the non-linear data using PLS and the residual compensator(RC) based on radial basis function network (RBFN). It compensates for the error of the non-linear data using the RC based on RBFN. The experimental result is given to verify its efficiency compared with those of previous works.