• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.358-360
• /
• 2006

LDA는 그룹간 간격을 최대화하고 그룹내 분산을 최소화하는 선형변환을 구함으로써 차원 감소된 공간에서 분별력(classification performance)을 높이는 선형 차원 감소 방법이다. 본 논문에서는 저샘플 문제(undersampled problem)에서 LDA를 적용할 수 있도록 QR-분해를 이용한 효율적인 차원 감소 방법을 제안한다. 특히 제안되는 방법은 문서 분류 문제에서처럼 한 문서가 몇 개의 카테고리에 중복적으로 속하는 경우 등 데이터의 독립성이 보장되지 않는 경우에도 효과적으로 적용될 수 있다는 장점이 있다.

• Journal of KIISE
• /
• v.45 no.1
• /
• pp.69-75
• /
• 2018

In the streaming data where data samples arrive sequentially in time, it is difficult to apply the dimension reduction method based on batch learning. Therefore an incremental dimension reduction method for the application to streaming data has been studied. In this paper, we propose an incremental linear discriminant analysis method using the least squared error solution. Instead of computing scatter matrices directly, the proposed method incrementally updates the projective direction for dimension reduction by using the information of a new incoming sample. The experimental results demonstrate that the proposed method is more efficient compared with previously proposed incremental dimension reduction methods.

• The KIPS Transactions:PartD
• /
• v.19D no.2
• /
• pp.147-150
• /
• 2012

While LDA is a supervised dimension reduction method which finds projective directions to maximize separability between classes, the performance of LDA is severely degraded when the number of labeled data is small. Recently semi-supervised dimension reduction methods have been proposed which utilize abundant unlabeled data and overcome the shortage of labeled data. However, matrix computation usually used in statistical dimension reduction methods becomes hindrance to make the utilization of a large number of unlabeled data difficult, and moreover too much information from unlabeled data may not so helpful compared to the increase of its processing time. In order to solve these problems, we propose an ensemble approach for semi-supervised dimension reduction. Extensive experimental results in text classification demonstrates the effectiveness of the proposed method.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10a
• /
• pp.70-72
• /
• 2001

기존의 CBF 기법은 데이타의 차원이 증가함에 따라 검색 성능이 급격히 저하되는 ‘Dimensional Curse’문제를 해결하기 위해 제안되었다. 그러나, 데이타의 양이 증가하고 차원이 증가할수록 검색 성능이 선형적인 감소를 보인다. 따라서, 본 논문에서는 CBF 기법의 성능 향상을 위해 멀티 디스크 환경을 기반으로 하는 병렬 CBF 기법을 제안한다. 제안하는 병렬 CBF 기법은 멀티 디스크 환경하에서 CBF가 지니는 특성을 이용하여 시그니쳐와 특징 벡터 데이타의 수평 분할 방법을 사용한다. 이를 통해, 제안하는 기법은 디스크 개수에 비례하여 선형적인 검색성능 향상을 가져온다.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.6 no.2
• /
• pp.157-163
• /
• 1994

An experimental study of deep-water breaking waves is performed by nonlinear wave evolution as well as superposition of different wave frequencies. Two-dimensional and three-dimensional wave instabilities and breakings are observed in nonlinear wave evolution. The wave energy evolves with almost the same initial wave energy before breaking but decreases significantly after breaking process. Large spilling and plunging waves are generated near e expected breaking location by means of faster waves overtaking slow waves at a certain point. More energy loss in vigorous plunging breakers is observed through breaking process.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.1
• /
• pp.90-95
• /
• 2009

This paper presents a new feature extraction algorithm which can deal with the problems of linear discriminant analysis, widely used for linear dimensionality reduction. The scatter matrices included in linear discriminant analysis are defined by the distances between each datum and its class mean, and those between class means and mean of whole data. Use of these scatter matrices can cause computational problems and the limitation on the number of features. In addition, these definition assumes that the data distribution is unimodal and normal, for the cases not satisfying this assumption the appropriate features are not achieved. In this paper we define a new scatter matrix which is based on the differently weighted distances between individual data, and presents a feature extraction algorithm using this scatter matrix. With this new method. the mentioned problems of linear discriminant analysis can be avoided, and the features appropriate for discriminating data can be achieved. The performance of this new method is shown by experiments.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.38 no.6
• /
• pp.657-664
• /
• 2001

It is possible to extract a 3D models from its multiple views using the self-calibration. Though it is possible to construct 3D models of objects from their multiple views, accuracy of 3D models depends on the fundamental matrix estimated between two views. In this paper, we show the fundamental matrix accuracy can be improved by taking a non-linear minimization technique. Furthermore, the corresponding points which are completely mismatches or have greater discrepancy errors in their locations, reduce the fundamental matrix accuracy. Thus, applying the Monte Carlo technique and the non-linear minimization Levenberg-Marquardt method to remove the outliers, we can estimate the fundamental matrix with the higher accuracy.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.4 no.3
• /
• pp.53-61
• /
• 1984

A nonlinear formulation of a beam finite element(NB6) on the total Lagrangian mode for the geometrically nonlinear analysis of two-dimensional elastic framed structures is presented. The NB6 beam element has been degenerated from the three-dimensional continuum by introducing the deep beam assumptions and consists of three reference nodes and three relative nodes. The element characteristics are derived by discretizing the beam equations of motion using the Galerkin weighted residual method and are reduced-integrated repeatedly for each loading step by the Newton-Raphson iteration techpique. Several numerical examples are given to demonstrate the accuracy and versatility of the proposed nonlinear NB6 beam element.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.105-105
• /
• 2021

본 연구에서는 2차원 비선형 천수모형과 수심평균된 스칼라 이송모형을 해석하는 수치모형에 대해 기술하였다. 수치모형의 정확성을 보장함과 동시에 안정성을 높이기 위해 유한체적법, 플럭스 재구성 및 minmod 제한자를 사용하였다. 비선형 천수방정식의 이송항과 바닥 경사항은 계산된 수심의 양수 보존과 흐름의 정상 상태를 보장하기 위한 second order well-balanced positivity preserving central-upwind method를 이용하여 수치적으로 이산화되었다. 마찬가지로, 이송-확산 방정식 내 이송항은 동일한 2차 풍상차분법을 통해 수치적으로 풀이하였다. 격자점 경계면에서의 불연속으로 인한 수치진동을 방지하기 위해 이송항의 계산에 포함된 보존항의 차이로 인해 발생하는 스칼라의 수치확산을 최소화하기 위해 무차원의 비소산함수를 도입하였다. 또한, 확산항은 유한차분법을 이용하여 이산화하였다. 제안된 수치모형은 시간미분항의 계산을 위해 오일러 기법을 적용하여 계산된 수심 및 스칼라의 양수 보존여부와 함께 정지된 흐름의 정상 상태의 보존여부를 확인하였다. 제안된 수치모형의 해석 정확성을 평가하기 위해 1, 2차원 공간 내 다양한 흐름 조건에서의 해석해를 이용한 3개의 벤치마크 테스트를 수행하였다. 평균 제곱근 오차(Root Mean Squared Error, RMSE)를 산정하여 수치모형의 성능을 정량적으로 평가하였으며, 비소산함수를 적용함에 따라 스칼라의 수치확산이 감소하게 되었음을 확인하였다. 또한, 세 차례의 벤치마크 테스트 결과는 공통적으로 수치모형에 의해 계산된 결과값이 비소산함수를 고려함에 따라 해석해와 잘 일치함을 확인하였다.

• Computational Structural Engineering
• /
• v.10 no.1
• /
• pp.185-191
• /
• 1997

Beam - and arch-like structures are two-dimensional bodies characterized by the fact of small thickness compared to the length of structures. Owing to this geometric feature, linear displacement approximations through the thickness such as Kirchhoff and Reissner-Mindlin theories which are more accessible one dimensional problems have been used. However, for accurate analysis of the behavior in the regions where the state of stresses is complex, two-dimensional linear elasicity or relatively high order of thickness polynomials is required. This paper analyses accuracy according to the order of thickness polynomials and introduces a technique for model combination for which several different polynomial orders are mixed in a single structure.