• 한국언어정보학회지:언어와정보
• /
• 제7권1호
• /
• pp.1-22
• /
• 2003

Quantifier floating is one of the much studied phenomena in natural languages where quantifying expressions may appear in places other than their original prenominal one. Its presence is especially prominent in languages such as Korean that allow more or less free word order. We find that, in addition to what is described in the literature, there are other remarkable regularities in the way the language allows quantifiers to “float” with respect to various constructions including coordination, relative clauses, and embedded clauses. These regularities are captured syntactically in a combinatory categorial grammar (CCG) framework for Korean. We also show how to derive semantic representations for Korean quantifier floating in the same CCG framework.

• 한국전자거래학회지
• /
• 제22권4호
• /
• pp.159-173
• /
• 2017

기술력 기반의 중소벤처기업에 대한 기술금융 지원을 위해 정부는 2014년 7월부터 기술보증기금 및 일정 자격을 갖춘 민간 기술신용평가사에게 일종의 기술력 등급평가인 기술신용평가를 실시하여 은행의 여신에 활용토록 하였다. 본 논문에서는 최근까지의 기술신용평가 현황 및 한국신용정보원에서 축적하고 있는 기술평가 관련 가용 지표들에 대한 선행 연구를 개략적으로 살펴본 후 기술평가등급점수에 유의적인 영향을 미치는 지표(indicator)를 통상적인 다중회귀기법으로 탐색할 것이다. 본 논문의 관심 대상인 지표 별 등급 영향도와 모형의 적합도는 대표적인 기계학습 분류기(classifier)인 일반화가속모형(Generalized Boosting Model; GBM)을 적용하여 분석하였는 바, 주요 지표를 독립변수(feature)로 투입하여 지표의 상대적 중요성 및 분류 정확도를 산출하였다. 분석결과 회귀모형과 기계학습 모형 간 지표별 상대적인 중요도는 크게 차이나지 않는 것으로 분석되었으나, GBM 모형의 경우 회귀모형에 비해서 이노비즈인증, 연구소 및 연구개발전담부서 보유, 특허등록건수, 벤처확인 지표 등 기술개발역량이 상대적으로 기술등급에 더 큰 영향을 미치는 것으로 분석되었다.

• 응용통계연구
• /
• 제18권1호
• /
• pp.43-56
• /
• 2005

본 연구에서는 AR(1) 과정을 따르는 시계열 모형에서 가산적 이상치(Additive Out-lier)가 존재하는 경우, 1차 자기상관계수에 대한 로버스트 추정방법으로 Rupport 와 Carroll (1980)에 의해 회귀모형에서 제안된 L-추정법 형태의 절사최소제곱추정 (PE 추정)방법을 제안하였다. 더불어 X축의 이상치에 대한 비중강하(down-weight)의 방법으로 Mallows의 가중함수를 고려한 유계영향 절사최소제곱 (bounded influence PE, BIPE)추정량을 제안하였으며 모의 실험을 통하여 각 추정량의 효율성을 비교하였다. 모의실험 결과, 다양한 자료의 오염률상에서 일반화 LAD추정치를 예비 추정치로 고려한 BIPE(LAD)-추정량의 효율이 좋은 것으로 나타났다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제14권4호
• /
• pp.276-287
• /
• 2014

Neural fuzzy system (NFS) is basically a fuzzy system that has been equipped with learning capability adapted from the learning idea used in neural networks. Due to their outstanding system modeling capability, NFS have been widely employed in various applications. In this article, we intend to discuss several ideas regarding the learning of NFS for modeling systems. The first issue discussed here is about structure learning techniques. Various ideas used in the literature are introduced and discussed. The second issue is about the use of recurrent networks in NFS to model dynamic systems. The discussion about the performance of such systems will be given. It can be found that such a delay feedback can only bring one order to the system not all possible order as claimed in the literature. Finally, the mechanisms and relative learning performance of with the use of the recursive least squares (RLS) algorithm are reported and discussed. The analyses will be on the effects of interactions among rules. Two kinds of systems are considered. They are the strict rules and generalized rules and have difference variances for membership functions. With those observations in our study, several suggestions regarding the use of the RLS algorithm in NFS are presented.

• 한국소음진동공학회논문집
• /
• 제15권1호
• /
• pp.20-28
• /
• 2005

이 논문에서는 기존의 6-축 혹은 9-축 머리진동 측정장치(bite-bar)를 이용한 머리진동 측정에서 직면하는 불확도 인자들에 대한 문제점을 우선 소개한다. 이들 불확도 인자들은 머리의 6-자유도 운동 성분의 추정에 필요한 모든 측정 요소들을 측정하지 않았다는 한계점에서 유발함을 보인다. 이러한 문제점을 극복하기 위하여 4개의 3축 가속도 센서로 구성된 새로운 머리진동 측정장치(12-axis bite-bar)의 모델을 제안한다. 본 모델은 측정 기준점에서의 선형 3축 가속도 뿐 아니라 3축 각 가속도와 함께 6 종의 2차 각속도 성분들의 추정 또한 가능하게 한다. 이러한 12 성분의 추정 모델로부터 비로소 머리의 임의 점에서 6-자유도 운동 성분의 계산이 가능함을 이론적으로 규명한다. 이러한 이론적 배경에 기반을 두고 설계 제작된 12-축 머리진동 장치(12-axis bite-bar)를 소개한다. 본 장치를 이용하여 얻어진 실험 결과 소개 뿐 아니라 기존의 측정장치의 측정 결과와 비교 분석 내용을 소개한다.

• 대한교통학회지
• /
• 제25권2호
• /
• pp.121-132
• /
• 2007

본 연구는 대중교통 통행배정을 위한 수단 및 경로선택의 기준으로서 도보시간, 대기시간(환승대기시간 포함), 차내시간, 환승시간(환승도보시간), 요금 및 각 요소별 가중치로 구성된 일반화비용 모형을 제시하였다. 모형의 정산을 위해 현실상황에 직면하도록 설계된 선호도 조사를 실시하여 자료를 수집/분석 하였으며, 한계대체율 및 임금율법을 적용하여 일반화비용 모형의 각 변수별 가중치를 도출하였다. 그 결과 서울시내간 통행의 경우 차내시간 대비 도보시간의 가중치는 1.507, 대기시간은 1.749, 환승시간은 1.474, 요금은 1.476이며, 서울경기간의 경우, 각각 1.827, 1.967, 1.015, 0.857로 도출되었다. 통계검정 결과 두 모형간에는 차이가 있는 것으로 나타났으며, 각 변수는 유의미 한 것으로 나타났다. 이 결과를 활용, 서울시 대중교통체계 개편 이후의 통행지표를 거시적 정량적으로 분석한 결과 서울시내간, 서울경기간 평균총통행시간은 30.23분, 63.29분으로 나타났으며, 일반화비용은 각각 2,510원, 3,880원으로 추정되었다.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2007년도 춘계학술대회논문집
• /
• pp.387-392
• /
• 2007

Turbomachinery such as turbines, pumps and compressors, which are installed in transportation systems such as warships, submarines and space vehicles, etc., often perform crucial missions and are exposed to potential dangerous impact environments such as base-transferred shock forces. To protect turbomachinery from excessive shock forces, it may be needed to accurately analyze transient responses of rotors, considering the dynamics of mount designs to be applied with. In this study a generalized FE transient response analysis model, introducing relative displacements, is firstly proposed to accurately predict transient responses of a flexible rotor-bearing system with mount systems to base-transferred shock forces. In the transient analyses the state-space Newmark method of a direct time integration scheme is utilized, which is based on the average velocity concept. Results show that for the identical mount systems considered, the proposed FE-based detailed flexible rotor model yields more reduced transient vibration responses to the same shocks than a conventional simple model or a Jeffcott rotor. Hence, in order to design a rotor-bearing system with a more compact light-weighted mount system, preparing against any potential excessive shock, the proposed FE transient response analysis model herein is recommended.

• 한국소음진동공학회논문집
• /
• 제17권12호
• /
• pp.1208-1216
• /
• 2007

Turbomachinery such as turbines, pumps and compressors, which are installed in transportation systems, including aircrafts, ships, and space vehicles, etc., often perform crucial missions and are exposed to potential dangerous impact environments such as base-transferred shock forces. To protect turbomachinery from excessive shock forces, it may be needed to accurately analyze transient responses of their rotors, considering the dynamics of mount designs to be applied. In this study a generalized FE transient response analysis model, introducing relative displacements, is proposed to accurately predict transient responses of a flexible rotor-bearing system with mount systems to base-transferred shock forces. In the transient analyses the state-space Newmark method of a direct time integration scheme is utilized, which is based on the average velocity concept. Results show that for the identical mount systems considered, the proposed FE-based detailed flexible rotor model yields more reduced transient vibration responses to the same shocks than a conventional simple model, obtained by treating a rotor as concentrated lumped mass, equivalent spring and a damper or Jeffcott rotor model. Hence, in order to design a rotor-bearing system with a more compact light-weighted mount system, preparing against any potential excessive shock, the proposed FE transient response analysis model herein is recommended.

• Structural Engineering and Mechanics
• /
• 제19권3호
• /
• pp.297-320
• /
• 2005

In recent years the pure displacement formulation for plate elements has not been as popular as other formulations. We revisit the pure displacement formulation for shear-deformable plate elements and propose a family of N-node, displacement-compatible, fully-integrated, pure-displacement, triangular, Mindlin plate elements, MIN-N. The development has been motivated by the relative simplicity of the pure displacement formulation and by the success of the existing 3-node plate element, MIN3. The formulation of MIN3 is generalized to obtain the MIN-N family, which possesses complete, fully compatible kinematic fields, in which the interpolation functions for transverse displacement are one degree higher than those for rotations. General element-level formulas for the thin-limit Kirchhoff constraints are developed. The 6-node, 18 degree-of-freedom element MIN6, with cubic displacement and quadratic rotations, is implemented and tested extensively. Numerical results show that MIN6 exhibits good performance for both static and dynamic analyses in the linear, elastic regime. The results illustrate that the fully-integrated MIN6 element has excellent performance in the thin limit, even for coarse meshes, and that it does not require shear relaxation.

• 제어로봇시스템학회논문지
• /
• 제10권3호
• /
• pp.205-215
• /
• 2004

The characteristic ratio assignment (CRA) method〔1〕 is new polynomial approach which allows to directly address the transient responses such as overshoot and speed of response time in time domain specifications. The method is based on the relationships between time response and characteristic ratios($\alpha_i$ ) and generalized time constant (T), which are defined in terms of coefficients of characteristic polynomial. However, even though the CRA can apply to developing a linear controller that meets good transient responses, there are still some fundamental questions to be explored. For the purpose of this, we have analyzed several sensitivities of a linear system with respect to the changes of coefficients itself and $\alpha_i$ of denominator polynomial. They are (i) the unnormalized root sensitivity : to determine how the poles change as $\alpha_i$ changes, and (ii) the function sensitivity to determine the sensitivity of step response to the change of o, and to analyze the sensitivity of frequency response as o, changes. As an other important result, it is shown that, under any fixed T and coefficient of the lowest order of s in denominator, the step response is dominantly affected merely by $\alpha_1, alpha_2 and alpha_3$ regardless of the order of denominator higher than 4. This means that the rest of the$\alpha_i$ s have little effect on the step response. These results provide some useful insight and background theory when we select $\alpha_i$ and T to compose a reference model, and in particular when we design a low order controllers such as PID controller.