• 지능정보연구
• /
• 제12권4호
• /
• pp.1-14
• /
• 2006

본 연구는 상품추천을 위해 필요한 고객 선호도 추정모형(Customer Preference Estimation Model)을 제안하고, 이러한 선호도 추정결과에 따른 선호도 정보를 이용하여 궁극적으로 상품추천의 성과를 제고시키기 위한 방법을 제시하였다. 즉, 제품에 대한 고객 선호 영향요인들과 고객 선호도와의 관계를 모형화 함으로써 고객 선호도를 보다 더 정확히 추정할 수 있는 새로운 선호도 추정모형을 제안하였다. 이 제안모형은 선호도 영향요인들의 상대적인 가중치를 선호도 최적화 학습을 통해 도출함으로써, 보다 정확한 선호도 측정을 가능하게 해 준다. 한편, 이 모형의 타당성을 검증하기 위해서 본 연구에서는 가상서점 고객들을 대상으로 고객 선호도 정보를 수집한 후, 본 제안모형을 적용했을 때의 협업 필터링의 추천성과와 사전가중치 부여방식인 기존 선호도 계산식을 이용했을 경우의 추천성과를 비교 분석하였다. 이에 대한 실증분석 결과는 본 연구에서 제안한 선호도 추정모형을 적용했을 때의 협업 필터링의 성과가 기존 선호도 계산방식을 적용했을 때의 협업 필터링의 성과보다 더 우수한 것으로 나타났다.

• Communications for Statistical Applications and Methods
• /
• 제29권2호
• /
• pp.239-250
• /
• 2022

In many applications, we frequently encounter correlated multiple outcomes measured on the same subject. Joint modeling of such multiple outcomes can improve efficiency of inference compared to independent modeling. For instance, in developmental toxicity studies, fetal weight and number of malformed pups are measured on the pregnant dams exposed to different levels of a toxic substance, in which the association between such outcomes should be taken into account in the model. The number of malformations may possibly have many zeros, which should be analyzed via zero-inflated count models. Motivated by applications in developmental toxicity studies, we propose a Bayesian joint modeling framework for continuous and count outcomes with excess zeros. In our model, zero-inflated Poisson (ZIP) regression model would be used to describe count data, and a subject-specific random effects would account for the correlation across the two outcomes. We implement a Bayesian approach using MCMC procedure with data augmentation method and adaptive rejection sampling. We apply our proposed model to dose-response analysis in a developmental toxicity study to estimate the benchmark dose in a risk assessment.

• Smart Structures and Systems
• /
• 제10권4_5호
• /
• pp.485-499
• /
• 2012

Mode shape expansion is useful in structural dynamic studies such as vibration based structural health monitoring; however most existing expansion methods can not consider the modelling errors in the finite element model and the measurement uncertainty in the modal properties identified from vibration data. This paper presents a reliable approach for expanding mode shapes with consideration of both the errors in analytical model and noise in measured modal data. The proposed approach takes the perturbed force as an unknown vector that contains the discrepancies in structural parameters between the analytical model and tested structure. A regularisation algorithm based on the Tikhonov solution incorporating the L-curve criterion is adopted to reduce the influence of measurement uncertainties and to produce smooth and optimised expansion estimates in the least squares sense. The Canton Tower benchmark problem established by the Hong Kong Polytechnic University is then utilised to demonstrate the applicability of the proposed expansion approach to the actual structure. The results from the benchmark problem studies show that the proposed approach can provide reliable predictions of mode shape expansion using only limited information on the operational modal data identified from the recorded ambient vibration measurements.

• 응용통계연구
• /
• 제25권3호
• /
• pp.389-401
• /
• 2012

This paper derives a benchmark dose(BMD) and its 95% lower confidence limit(BMDL) using a semi-parametric regression model for small lead based changes in attention-deficit hyperactivity disorder(ADHD) scores in the first wave of the Children's Health and Environment Research(CHEER) survey data, which have been regularly collected in South Korea since 2005. Ha et al. (2009) showed that the appearance of ADHD symptoms had a borderline trend of increasing with the blood lead concentration. Butdz-J${\o}$rgensen (EFSA, 2010a) derived the BMDL of lead corresponding to a benchmark region of 1 full intelligent quotient (IQ) score using the raw data in Lanphear et al. (2005, EHP). European Food Safety Authority (EFSA, 2010b) determined the BMDL of $1.2{\mu}g/dl$ as a reference point for the characterization of lead when assessing the risk of the intellectual deficit measured by IQ scores. Kim et al. (2011) indicated that an even lower BMDL could be obtained based on the ADHD score; however, the BMDLs depended heavily upon the model assumptions. We show in this paper that a semi-parametric approach resolves the model dependence of BMDLs.

• 전자공학회논문지
• /
• 제51권1호
• /
• pp.202-208
• /
• 2014

로봇 시스템에 작용하는 마찰력은 비선형 형태이며 특히 저속에서의 정밀 제어를 어렵게 만든다. 본 논문에서는 로봇 연구에서 자주 활용되고 있는 벤치마크 로봇 시스템의 단축 모델에 대한 동적 마찰 보상 문제를 다룬다. 마찰 모델은 동적 특성 및 다양한 효과를 나타낼 수 있도록 비선형 동적 모델인 LuGre 모델을 고려한다. 본 논문에서 제안하는 제어기는 두 부분으로 구성된다. 동적 마찰의 추정 및 보상을 위해 Dual 관측기 기반의 적응 제어기를 사용한다. 마찰 추정 오차 및 나머지 외란을 보상하기 위해 PI 관측기를 추가로 설계한다. 모의실험을 통해 비선형 동적 마찰이 벤치마크 로봇 시스템의 제어 성능에 미치는 영향을 확인하고 제안된 제어기를 사용함으로써 동적 마찰에 대한 제어 성능이 향상됨을 보인다.

• Structural Engineering and Mechanics
• /
• 제43권3호
• /
• pp.371-394
• /
• 2012

Accurate finite element (FE) models are needed in many applications of Civil Engineering such as health monitoring, damage detection, structural control, structural evaluation and assessment. Model accuracy depends on both the model structure (the form of the equations) and the model parameters (the coefficients of the equations), and can be generally improved through that process of experimental reconciliation known as model updating. However, modelling errors, including (i) errors in the model structure and (ii) errors in parameters excluded from adjustment, may bias the solution, leading to an updated model which replicates measurements but lacks physical meaning. In this paper, an application of ambient-vibration-based model updating to a large-scale benchmark prototype of a building structure is reported in which both types of error are met. The error in the model structure, originating from unmodelled secondary structural elements unexpectedly working as resonant appendages, is faced through a reduction of the experimental modal model. The error in the model parameters, due to the inevitable constraints imposed on parameters to avoid ill-conditioning and under-determinacy, is faced through a multi-model parameterization approach consisting in the generation and solution of a multitude of models, each characterized by a different set of updating parameters. Results show that modelling errors may significantly impair updating even in the case of seemingly simple systems and that multi-model reasoning, supported by physical insight, may effectively improve the accuracy and robustness of calibration.

• Nuclear Engineering and Technology
• /
• 제55권6호
• /
• pp.2230-2245
• /
• 2023

KANT (KAIST Advanced Nuclear Tachygraphy) is a PWR core simulator recently developed at Korea Advance Institute of Science and Technology, which solves three-dimensional steady-state and transient multigroup neutron diffusion equations under Cartesian geometries alongside the incorporation of thermal-hydraulics feedback effect for multi-physics calculation. It utilizes the standard Nodal Expansion Method (NEM) accelerated with various Coarse Mesh Finite Difference (CMFD) methods for neutronics calculation. For thermal-hydraulics (TH) calculation, a single-phase flow model and a one-dimensional cylindrical fuel rod heat conduction model are employed. The time-dependent neutronics and TH calculations are numerically solved through an implicit Euler scheme, where a detailed coupling strategy is presented in this paper alongside a description of nodal equivalence, macroscopic depletion, and pin power reconstruction. For validation of the steady, transient, and depletion calculation with pin power reconstruction capacity of KANT, solutions for various benchmark problems are presented. The IAEA 3-D PWR and 4-group KOEBERG problems were considered for the steady-state reactor benchmark problem. For transient calculations, LMW (Lagenbuch, Maurer and Werner) LWR and NEACRP 3-D PWR benchmarks were solved, where the latter problem includes thermal-hydraulics feedback. For macroscopic depletion with pin power reconstruction, a small PWR problem modified with KAIST benchmark model was solved. For validation of the multi-physics analysis capability of KANT concerning large-sized PWRs, the BEAVRS Cycle1 benchmark has been considered. It was found that KANT solutions are accurate and consistent compared to other published works.

• 한국통신학회논문지
• /
• 제14권1호
• /
• pp.58-64
• /
• 1989

본 논문은 benchmark program의 명령어 mix들과 이 명령어 mix들의 분포와 빈도돌을 parameter로 해서 컴퓨터 성능평가를 행할 수 있는 확률적 모델을 제시한다. 이 모델을 Intel 8086/8083마이크로 프로세서의 성능 평가에 적용시켜 봄으로써, 현존하는 컴퓨터시스템뿐만 아니라 미개발 시스템을 위한 성능 평가 모델로도 가능하다는 것을 보였다.

• 한국원자력학회:학술대회논문집
• /
• 한국원자력학회 1997년도 추계학술발표회논문집(1)
• /
• pp.191-196
• /
• 1997

As a part of International Criticality Safety Benchmark Evaluation Project (ICSBEP), SAXTON critical experiments were reevaluated. The effects on $K_{eff}$ of the uncertainties in experiment parameters, fuel rod characterization, soluble boron, critical water level, core structure, $^{241}$ Am and $^{241}$ Pu isotope number densities, random pitch error, duplicated experiment, axial fuel position, model simplification, etc., were evaluated and added in benchmark-model $k_{eff}$. In addition to detailed model, the simplified model for Saxton critical experiments was constructed by omitting the top, middle, and bottom grids and ignoring the fuel above water.r.r.

• 터널과지하공간
• /
• 제30권6호
• /
• pp.573-590
• /
• 2020

본 논문에서는 국제공동연구인 DECOVALEX-2023 프로젝트 Task G의 연구 현황과 현재까지 수행된 benchmark 해석 결과를 소개하였다. Task G의 명칭은 'Safety ImplicAtions of Fluid Flow, Shear, Thermal and Reaction Processes within Crystalline Rock Fracture NETworks(SAFENET)'로, 결정질 암반 내 균열의 생성과 성장 메커니즘 및 균열에서 발생하는 열-수리-역학적 복합거동을 해석하기 위한 수치해석기법을 개발하는 데에 목표가 있다. Task G의 첫 번째 연구 테마는 결정질 암석 내 단일 균열의 역학적 거동에 대한 해석해(analytical solution)를 바탕으로 각 연구팀의 수치모델링기법을 개발 및 검증하는 Benchmark 해석이다. 본 연구에서는 3차원 입자기반 개별요소모델을 이용하여 단일 균열을 포함한 암석의 역학적 거동 특성을 모델링하고자 하였다. 이 모델에서는 상호독립적으로 거동하는 개별입자의 집합체를 통해 암석의 구조적 특징을 모사하고, 입자와 입자간 접촉에서 발생하는 역학적 거동을 개별요소해석모델인 3DEC을 통해 계산하게 된다. 해석 결과, 도메인의 경계응력으로 인해 균열에 유도되는 수직응력과 전단응력 수준은 변위 구속과 응력 재배치로 인해 이론적인 수치보다 낮게 나타났다. 그러나 수치모델에서 계산된 수직변위와 전단변위는 실제 균열의 유도 응력을 통해 추정된 해석해와 비교할 때 상당히 유사한 결과를 보였으며 균열의 응력-변위 관계를 합리적으로 재현할 수 있음을 확인하였다. 본 연구의 해석모델은 Task G에 참여하는 국외 연구팀들과의 의견 교류와 워크숍을 통해 지속적으로 개선하는 한편, 향후 다양한 조건의 실내시험에 적용하여 타당성을 검증할 예정이다.