• 대한전기학회논문지:전력기술부문A
• /
• 제48권2호
• /
• pp.82-88
• /
• 1999

In this paper, we propose a new structure optimization method of multilayer neural networks which begin and carry out learning from a bigger network. This method redundant links and neurons according to the rough set theory. In order to find redundant links, we analyze the variations of all weights and output errors in every step of the learning process, and then make the decision table from their variation of weights and output errors. We can find the redundant links from the initial structure by analyzing the decision table using the rough set theory. This enables us to build a structure as compact as possible, and also enables mapping between input and output. We show the validity and effectiveness of the proposed algorithm by applying it to the XOR problem.

• Journal of the Korean Statistical Society
• /
• 제27권1호
• /
• pp.25-55
• /
• 1998

Maximum likelihood estimates (MLEs) for recursive models of categorical variables are discussed under an EM framework. Since MLEs by EM often depend on the choice of the initial values for MLEs, we explore reasonable rules for selecting the initial values for EM. Simulation results strongly support the proposed rules.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1988년도 전기.전자공학 학술대회 논문집
• /
• pp.93-97
• /
• 1988

The sliding mode control is an effective method to establish robustness against parameter variations and disturbance. But, in sliding mode strategy, the control function is discontinuous on the hyperplane. Consequently, the control input chatters at high frequency. When we apply such a control to the induction motor drive system, that causes a torque ripple and current harmonics, which are harmful to the system. In this paper, we introduce a low pass filter between sliding mode control output and driver input to overcome that problem. The band-width of this filter is varied according to the error funtion to improve transient responses.

• 대한전기학회논문지:전기기기및에너지변환시스템부문B
• /
• 제48권4호
• /
• pp.203-211
• /
• 1999

A control approach for the robust position control of induction motors based on the binary disturbance observer is described. The conventional binary disturbance observer is used to remove the chattering problem of a sliding mode disturbance observer. However, the steady state error may exist in the conventional binary disturbance observer because it estimates external disturbance with a constant boundary layer. In order to overcome this problem, new binary disturbance observer with an integral augmented switching hyperplane is proposed. The robustness is achieved, and the continuous control is realized by employing the proposed observer without the chattering problem and the steady state error. The effectiveness of the proposed observer is confirmed by the comparative experimental results.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2003년도 춘계학술대회논문집
• /
• pp.932-937
• /
• 2003

When oil pressure of damper is lost or relative stiffness of spring drops in vibration system, it can be fatally dangerous situation. A fault diagnosis method for vibration system using Support Vector Machine(SVM)is suggested in the paper. SVM is used to classify input data or applied to function regression. System status can be classified by judging input data based on optimal separable hyperplane obtained using SVM which learns normal and abnormal status. It is learned from the relationship of system state variables in term of spring, mass and damper. Normal and abnormal status are learned using phase plane as in put space, then the learned SVM is used to construct algorithm to predict the system status quantitatively

• 대한전기학회논문지
• /
• 제38권9호
• /
• pp.671-683
• /
• 1989

A controller of interconnected power systems is investigated using an optimal multidimensional variable structure control. The switching hyperplane of the variable structure stabilizer is obtained by minimizing a quadratic performance index in continuous-time. A special feature of the optimal multidimensional variable structure stabilizer is that, when it is operated in the so-called sliding mode, the system response becomes insensitive to changes in the plant parameters. A digital simulation is performed by a digital computer using the Advanced Continuous Simulation Language (ACSL) package, which shows that the dynamic performance of the power system in response to mechanical torque changes is improved when optimal multidimensional variable sturcture stabilizers are employed.

• 대한전기학회논문지
• /
• 제39권1호
• /
• pp.49-56
• /
• 1990

An induction motor position control system based on the sliding mode control is presented. In the sliding mode control, control function is discontinuous on the hyperplane, which causes harmful effects such a s current harmonics and acoustic noise in the motor drive application. In this study, a low pass filter is introduced between the sliding mode controller output and the motor controller input to reduce these effects. The filter, however, makes the torque response slggish and the system performance may become poor in cost of chattering reduction. To overcome these problems, the bandwidth of the filer is varied according to the error function. It is shown that the proposed sliding mode control with variable-bandwidth filter shows good performance, which is confirmed through experiments.

• 대한수학회지
• /
• 제58권1호
• /
• pp.109-122
• /
• 2021

In this paper, we consider the problem of finding the hypersurface Mn in the Euclidean (n + 1)-space ℝn+1 that satisfies an equation of mean curvature type, called singular minimal hypersurface equation. Such an equation physically characterizes the surfaces in the upper half-space ℝ+3 (u) with lowest gravity center, for a fixed unit vector u ∈ ℝ3. We first state that a singular minimal cylinder Mn in ℝn+1 is either a hyperplane or a α-catenary cylinder. It is also shown that this result remains true when Mn is a translation hypersurface and u is a horizantal vector. As a further application, we prove that a singular minimal translation graph in ℝ3 of the form z = f(x) + g(y + cx), c ∈ ℝ - {0}, with respect to a certain horizantal vector u is either a plane or a α-catenary cylinder.

• Communications for Statistical Applications and Methods
• /
• 제30권3호
• /
• pp.245-258
• /
• 2023

This study proposes a modification to the objective function of the support vector machine for the linearly non-separable case of a binary classifier y_i ∈ {-1, 1}. The modification takes into account the position of each data item x_i from its corresponding class centroid. The resulting optimization function involves the centroid mean vector, and the spread of data besides the support vectors, which should be minimized by the choice of hyper-plane β. Theoretical assumptions have been tested to derive an optimal separable hyperplane that yields the minimal misclassification rate. The proposed method has been evaluated using simulation studies and real-life COVID-19 patient outcome hospitalization data. Results show that the proposed method performs better than the classical linear SVM classifier as the sample size increases and is preferred in the presence of correlations among predictors as well as among extreme values.

• 경영정보학연구
• /
• 제24권4호
• /
• pp.23-40
• /
• 2022

Support Vector Machine(SVM)은 기업부실 예측문제 등 다양한 분야에서 성공적으로 활용되어 왔으나 범주 불균형 문제가 존재하는 경우 다수 범주의 경계영역은 확장되는 반면, 소수 범주의 경계영역은 축소되고 분류 경계선이 소수 범주로 편향되어 분류 성과에 부정적인 영향을 미치는 것으로 보고되고 있다. 본 연구는 범주 불균형 문제에 대한 대칭 마진 SVM(EMSVM)의 한계점을 개선하기 위하여 비대칭 마진 SVM(UMSVM)과 임계점 이동 기법을 결합한 최적화 비대칭 마진 SVM인 OPT-UMSVM을 제안한다. OPT-UMSVM은 소수 범주 방향으로 치우진 분류 경계선을 다수 범주로 재이동함으로써 소수 범주의 민감도를 개선하고 최적화된 분류 성과를 산출함으로써 SVM의 일반화 능력을 향상시키는 장점을 가진다. OPT-UMSVM의 성과 개선 효과를 검증하기 위하여 불균형 비율이 상이한 5개의 표본군을 구성하여 10-fold 교차타당성 검증을 수행한 결과는 다음과 같다. 첫째, 범주 불균형이 미미한 표본에서 UMSVM은 EMSVM의 성과 개선 효과가 미약한 반면, 범주 불균형이 심화된 표본에서 UMSVM은 EMSVM의 성과개선에 크게 공헌하고 있다. 둘째, OPT-UMSVM은 EMSVM 및 기존의 UMSVM과 비교하여 범주 균형 및 범주 불균형 표본 모두에서 보다 우수한 성과를 가지고 있으며, 특히 범주 불균형이 심화된 표본에서 유의적인 성과 차이를 보였다.