• Journal of Advanced Marine Engineering and Technology
• /
• v.35 no.2
• /
• pp.271-277
• /
• 2011

An identification method of the structural damping matrix for a steel bar by the genetic algorithm is proposed. Supposing the damping matrix were in proportion to the stiffness matrix, the proportional factors can be identified from the curve fitting of the experimental frequency response function(FRF) by the genetic algorithm. Applying the identified damping matrix to FEM of a beam model, the values of the objective function could be reduced to about 1/60 in comparison with conventional FEM model without damping. The damping matrices of some sub-structures which have large damping partly could be identified by the algorithm, and they could be used as some parts of the FEM model for a whole structure.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.10a
• /
• pp.135-138
• /
• 2007

Essay education has been multifariously accomplished through school curriculums, private institutions and online education systems. However we can easily find misspellings from students' writing, and them still not being able to describe their own experiences logically and skillfully in schools. In this paper, we propose the web-based essay correction system using mash-up service, ECSUM, implemented for improvement of a students' logical writing ability applying a logical writing correction module with six steps. Using the proposed system, students can obtain logical writing skills using a mash-up service and teachers can help their students to improve their logical ability of self-expression.

• Journal of Information Management
• /
• v.19 no.1
• /
• pp.95-122
• /
• 1988

From the lack of common rule for the description of the characters and terms in bibliographic databases, it was hard to share information with other organizations and to obtain relevant bits of information through online retrieval. In this paper, the authors suggest a rule for the transcription of symbols and letters, that are found in articles. but impossible to input through CRT terminals, into the symbols and letters which is capable of input and retrieval. And, in the 'Hangul orthography' and 'Description rule of the borrowed words' which are officially announced by the Ministry of Education, more than two ways are permited for the expression of terms. In that case, to improve retrieval efficiency and to prevent 'confusion in description, they are regulated.

• Journal of the Korean Magnetics Society
• /
• v.8 no.2
• /
• pp.49-56
• /
• 1998

Explicit and closed-form expressions describing magnetic behaviors of uniaxial magnetic materials are derived. Explicit and closed-form expressions for magnetic torque, magnetization orientation, and thier derivative functions with respect to the intensity and the orientation of an applied field. Those explicit expressions could give elegant mehtods of measuring magnetic anisotropy and saturation magnetization by the least squre fitting. In addition, on could use them to study the distribution function of magnetic propeties of a non-interacting magnetic aggregate.

• Annual Conference on Human and Language Technology
• /
• 2004.10d
• /
• pp.82-89
• /
• 2004

본 논문에서는 한국어 의학 논문의 요약문으로부터 자동 영문 MeSH 키워드 제안 시스템을 소개하고, 띄어쓰기 변이(spacing variant) 문제를 해결할 수 있는 방법을 제안한다. 띄어쓰기 변이란 표준 한글 맞춤법에 비해 다르게 띄어쓰기된 것을 말한다. 이를 위해 시소러스에는 생성 가능한 모든 띄어쓰기 변이 대신에 최대 띄어쓰기 어구만을 저장하고, 문서에서 K-MeSH 용어를 찾기 위해 음절단위 부분문자열 검색을 사용한다. 이 방법으로 한국어 의학 논문의 요약문에서 K-MeSH 용어를 추출한 후, TF-IDF 순위 함수를 이용하여 상위 10위내의 키워드를 저자가 선정한 영문 키워드와 비교한 결과 58%가 일치하였다. 이는 기존 방법에 비해 42%정도의 시소러스 크기가 축소되었고, 상위 10위내에서 영문 MeSH 키워드 추천 재현률이 약 7.8% 증가한 것으로 효과적인 방법임을 보여주었다.

• Journal of Ocean Engineering and Technology
• /
• v.22 no.3
• /
• pp.41-46
• /
• 2008

It is a recent trend for advanced ships and submarines to incorporate composite structures with viscoelastically damping material. Much research has been done on curve-fitting techniquesto identify vibration characteristic parameters such as natural frequencies, modal damping ratios, and mode shapes of the composite structure. In this study, an advanced technique for accurately determining vibration characteristic of a circular cylindrical shell-attached viscoelastically damping material is used, based on a multi-degree of freedom (MDOF) curve-fitting method. First, an initial value is obtained by using a linear least square method. Next, using the initial value, the exact modal parameters of the composite circular cylindrical shell are obtained by using a nonlinear least square method. Results show computation time is greatly decreased and accurate results are obtained by the MDOF curve-fitting method.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.273-275
• /
• 2016

외국인 유학생과 국내 체류 외국인을 포함하여 한국어를 학습하고자 하는 외국인이 지속적으로 증가함에 따라, 외국인 한국어 학습자의 교육에 대한 관심도 높아지고 있다. 기존 맞춤법 검사기는 한국어를 충분히 이해할 수 있는 한국인의 사용에 중점을 두고 있어, 외국인 한국어 학습자가 사용하기에는 다소 부적절하다. 본 논문에서는 한국어의 문맥 특성과 외국인의 작문 특성을 반영한 한국어 교정 방식을 제안한다. 제안하는 시스템에서는 말뭉치에서 추출한 어절 바이그램에 대한 음절 역색인을 구성하여 추천 표현을 빠르게 제시할 수 있으며, 키보드 후킹에 기반한 사용자인터페이스를 제공하여 사용자 편의를 높인다.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.20 no.10
• /
• pp.946-952
• /
• 2010

A method for the identification of structural characteristic parameters of a steel bar in the matrices form such as stiffness matrices and mass matrices from frequency response function(FRF) by genetic algorithm is proposed. As the method is based on the finite element method(FEM), the obtained matrices have perfect physical meanings if the FRFs got from the analysis and the FRFs from the experiments were well coincident each other. The identified characteristic matrices from the FRFs with maximun 40 % of random errors by the genetic algorithm are coincident with the characteristic matrices from exact FEM FRFs well each other. The fitted element diameters by using only 2 points experimental FRFs are similar to the actual diameters of the bar. The fitted FRFs are good accordance with the experimental FRFs on the graphs. FRFs of the rest 9 points not used for calculating could be fitted even well.

• Annual Conference of KIPS
• /
• 2019.05a
• /
• pp.447-450
• /
• 2019

본 논문은 한국어 문장의 감성 분류를 위해 문장의 형태소, 음절, 자소를 입력으로 하는 합성곱층과 DenseNet 을 적용한 Text Multi-channel DenseNet 모델을 제안한다. 맞춤법 오류, 음소나 음절의 축약과 탈락, 은어나 비속어의 남용, 의태어 사용 등 문법적 규칙에 어긋나는 다양한 표현으로 인해 단어 기반 CNN 으로 추출 할 수 없는 특징들을 음절이나 자소에서 추출 할 수 있다. 한국어 감성분석에 형태소 기반 CNN 이 많이 쓰이고 있으나, 본 논문에서 제안한 Text Multi-channel DenseNet 모델은 형태소, 음절, 자소를 동시에 고려하고, DenseNet 에 정보를 밀집 전달하여 문장의 감성 분류의 정확도를 개선하였다. 네이버 영화 리뷰 데이터를 대상으로 실험한 결과 제안 모델은 85.96%의 정확도를 보여 Multi-channel CNN 에 비해 1.45% 더 정확하게 문장의 감성을 분류하였다.

• Journal of Convergence for Information Technology
• /
• v.9 no.9
• /
• pp.158-165
• /
• 2019

Through the use of speech software technology, this paper examines the effects of voice interactions within the inner circle of English. The fundamental frequency (F0) was obtained by analyzing native speakers (aged 30-55) speech effects based on nationality, age, and gender. The findings within this study reveal that the Caucasian British female (age 33) and the Caucasian American male (age 55) produced the most interactive speech. The contributing factor is the students' experience with various language styles throughout their language acquisition studies. The results of this study are compatible with $Traunm{\ddot{u}}eller$ & Eriksson (1995) and previous studies which agree that continuous speech above average is paramount towards student engagement and interactions.