• The Korean Journal of Applied Statistics
• /
• v.28 no.6
• /
• pp.1209-1216
• /
• 2015

Competing-risks events are often observed in a clustered clinical study such as a multi-center clinical trial. We propose a joint modelling approach via a shared frailty term for competing risks survival data from a cluster. For the inference we use the hierarchical likelihood (or h-likelihood), which avoids an intractable integration. We derive the corresponding h-likelihood procedure. The proposed method is illustrated via the analysis of a practical data set.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2017-2019
• /
• 2003

열악한 환경에서 동작해야 하거나 물리적 접근이 어려운 곳에 장착되는 센서 시스템의 경우, 유선에 의한 정보전달이 어려울 뿐만 아니라 센서 내 전원설비가 제한적인 수도 있다. 따라서 본 논문에서는 이러한 문제점에 대한 해결책으로서 밧데리없이 유도결합에 의하여 원격 센서로부터 정보 취득이 가능한 한 방법을 제안하였다. 이 방법은 전원공급에 의한 유도 결합식의 원격 센서 시스템과는 달리, 원격 센서의 정전용량을 변 복조 과정 없이 재귀적 최소 자승 추정법에 의해 센서의 정전용량을 고정도로 추정하는 것이다. 이를 위하여 시 스템의 유도결합 모델을 사용하였으며, 모델 파라미터가 시간에 따라 변화하는 경우에는 망각률을 사용하여 정확도가 높은 원격 센서 시스템을 구현할 수 있었다.

• Journal of the Korea Computer Industry Society
• /
• v.5 no.3
• /
• pp.377-382
• /
• 2004

An unconstrained three-dimensional cutting problem describes the process of finding the cutting pattern that yields the maximum total profit-sum for the small parallelepipeds pieces cut from a large parallelepiped box and there is no limit to the number of pieces to be cut. The problem is a classic NP-hard. The bottom-up approach, which generates all of the feasible cutting patterns by combining two other cutting patterns, can be applied to the three-dimensional problem. We introduce a build and new branching strategies for the unconstrained three-dimensional cutting problem. The strategies are all generalized from the branching strategies proposed by G et at. to solve unconstrained two-dimensional cutting problems.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.245-248
• /
• 2007

서비스 지향 아키텍처(SOA)는 기업이 적정한 비용으로 오늘날 가장 중요한 IT 요구사항인 민첩성과 유연성을 만족시킬 수 있는 IT 아키텍처 수립의 토대로, 경영환경이 빠르게 급변하는 최근에 떠오른 이슈이다. 기존의 서비스 지향 아키텍처의 서비스 개발은 주로 비즈니스 환경에서 어플리케이션방향으로 접근하는 하향식 방법이 대부분이었으나, 본 논문에서는 이미 개발된 컴포넌트 기반 시스템에서 접근하는 상향식 서비스 식별법을 제안한다. 본 논문에서 제안하는 방법은 사용자의 이벤트 정보를 담고 있는 GUI Flow-Event 를 이용하여 컴포넌트를 서비스로 식별하고, 서비스 식별시 발생하는 서비스간 연관관계에 따른 문제점을 최소화함으로써 비즈니스 도메인에 더욱 가깝고, 약결합을 지향하는 서비스로 식별할 수 있다.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.55-71
• /
• 2013

The prediction model is the main factor affecting the performance of a knowledge-based system for bankruptcy prediction. Earlier studies on prediction modeling have focused on the building of a single best model using statistical and artificial intelligence techniques. However, since the mid-1980s, integration of multiple techniques (hybrid techniques) and, by extension, combinations of the outputs of several models (ensemble techniques) have, according to the experimental results, generally outperformed individual models. An ensemble is a technique that constructs a set of multiple models, combines their outputs, and produces one final prediction. The way in which the outputs of ensemble members are combined is one of the important issues affecting prediction accuracy. A variety of combination schemes have been proposed in order to improve prediction performance in ensembles. Each combination scheme has advantages and limitations, and can be influenced by domain and circumstance. Accordingly, decisions on the most appropriate combination scheme in a given domain and contingency are very difficult. This paper proposes a confidence-based selection approach as part of an ensemble bankruptcy-prediction scheme that can measure unified confidence, even if ensemble members produce different types of continuous-valued outputs. The present experimental results show that when varying the number of models to combine, according to the creation type of ensemble members, the proposed combination method offers the best performance in the ensemble having the largest number of models, even when compared with the methods most often employed in bankruptcy prediction.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.1
• /
• pp.62-68
• /
• 2009

Two new generation gap models with modified parent-centric recombination(PCX) operator are proposed. First, the self-adaptation generation gap(SGG) model is a control method that keeps a replaced probability of parents by offspring to a certain level which obtains better performance. Second, virtual cluster generation gap(VCGG) is provided to extend distances among parents using clustering, which causes it to diversify individuals. In this model, distances among parents can be controlled by size of clusters. To demonstrate the effectiveness of our two proposed approaches, experiments for three standard test problems are executed and compared to most competing current approaches, CMA-ES and Generalized Generation Gap(G3) with PCX. It is shown two proposed methods are superior to consistently other approaches in the study.

• Korean Chemical Engineering Research
• /
• v.60 no.3
• /
• pp.398-406
• /
• 2022

The outer membrane of Gram-negative bacteria is the outermost layer of cellular environment in which numerous biophysical and biochemical processes are in action sustaining viability. Advances in cell engineering enable modification of bacterial genetic information that subsequently alters membrane physiology to adapt bacteria to specific purposes. Surface display of a functional molecule on the outer membranes is one of strategies that directs host cells to respond to a specific extracellular matter or stimulus. While intracellular expression of a functional peptide or protein fused to a membrane-anchoring motif is commonly practiced for surface display, the method is not readily applicable to exogenous or large proteins inexpressible in bacteria. Chemical conjugation at reactive groups naturally occurring on the membrane might be an alternative, but often compromises fitness due to non-specific modification of essential components. Herein, we demonstrated two distinct approaches that enable site-specific decoration of the outer membrane with a fluorescent agent in Escherichia coli. An unnatural amino acid genetically incorporated in a surface-exposed peptide could act as a chemoselective handle for bioorthogonal dye labeling. A surface-displayed α-helical domain originating from a part of a selected heterodimeric coiled-coil complex could recruit and anchor a green fluorescent protein tagged with a complementary α-helical domain to the membrane surface in a site- and hetero-specific manner. These methods hold a promise as on-demand tools to confer new functionalities on the bacterial membranes.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.958-961
• /
• 2019

Convolutional neural network(CNN)의 등장으로 얼굴 표현 인식 연구는 많은 발전을 이루었다. 그러나, 기존의 CNN 접근법은 미리 학습된 훈련모델에서 Multiple-level 의 의미적 맥락을 포함하지 않는 Attention-embedded 문제가 발생한다. 사람의 얼굴 감정은 다양한 근육의 움직임과 결합에 기초하여 관찰되며, CNN 에서 딥 레이어의 산출물로 나온 특징들의 결합은 많은 서브샘플링 단계를 통해서 class 구별와 같은 의미 정보의 손실이 일어나기 때문에 전이 학습을 통한 올바른 훈련 모델 생성이 어렵다는 단점이 있다. 따라서, 본 논문은 Backbone 네트워크의 Multi-level 특성에서 Channel-wise Attention 통합 및 의미 정보를 포함하여 높은 인식 성능을 달성하는 Densely Backwarnd Attention(DBA) CNN 방법을 제안한다. 제안하는 기법은 High-level 기능에서 채널 간 시멘틱 정보를 활용하여 세분화된 시멘틱 정보를 Low-level 버전에서 다시 재조정한다. 그런 다음, 중요한 얼굴 표정의 묘사를 분명하게 포함시키기 위해서 multi-level 데이터를 통합하는 단계를 추가로 실행한다. 실험을 통해, 제안된 접근방법이 정확도 79.37%를 달성 하여 제안 기술이 효율성이 있음을 증명하였다.

• Tunnel and Underground Space
• /
• v.14 no.4
• /
• pp.248-260
• /
• 2004

Damage in brittle rock due to stress increase starts from initiation of microcracks, and then results in failure by forming macro failure planes due to propagation and coalescence of these discrete cracks. Conventionally, continuum approaches using macro-failure criteria or a number of elasto-plastic models have been major solution to implement rock damage and failure. However, actual brittle failure processes can be better described in phenomenological approach if initiation and propagation of discrete fractures are explicitly considered. This study presents damage and failure process of rock using a boundary element code, FRACOD, which has been developed to model fracturing process of rocks. Through a series of numerical uniaxial compressive tests, the feasibility of the developed model was verified, and realistic rock failure process was reproduced considering scale effects in rocks. In addition, the fracturing process and the corresponding rock damage in the vicinity of deep shaft in rock mass were presented as an application of this approach. This approach will be expected to contribute to finding better engineering solutions for the analysis of stability problems in brittle rock masses.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.519-522
• /
• 2004

본 논문에서는 외향 기반 접근법을 기반으로 한 로봇의 위치 추정 알고리즘을 제안한다. 로봇이 작업을 수행할 공간에서 강한 상관관계를 갖는 영상들을 취득하여 eigenspace로 투영 시킴으로써 주성분의 추출을 수행한다. 이 추출된 주성분은 신경 회로망을 이용해 eigenspace에서의 연속 외향 함수(continuous appearance function)로 나타낼 수 있다. 로봇의 위치 추정을 위해 새로운 영상이 주어지면 이것을 eigenspace로 투영 시킨 후 연속 외향 함수를 통해 로봇의 현재 위치를 추정한다. 최종적으로는, 영상안의 데이터에 칼만 필터를 적용함으로써 로봇의 정확한 위치와 영상으로 획득된 정보 사이의 오차를 이용하여 보다 정확한 이동 로봇의 위치를 추정하는 알고리즘을 제안한다.