• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2
• /
• pp.101-104
• /
• 1996

The pressure field for the Pekeris model which is consisted of a surface layer of fresh water overlying a thick (essentially semi-infinite) bottom layer of coarse sand is investigated experimentally in a anechoic tank scaled down 1/1,000. The water depth and frequency have been controlled so that the continuous mode integrand has not a resonance close to cut-off, and the experimental results relatively well agree with the theory which is represented as the sum of the discrete modes over a range about 10 times the water depth.

• International Journal of Air-Conditioning and Refrigeration
• /
• v.6
• /
• pp.14-26
• /
• 1998

This paper deals with approximate analytical solutions for the two-region one-dimensional model describing the charging process of stratified thermal storage tanks at variable inlet temperature with momentum-induced mixing. An arbitrarily increasing inlet temperature is decomposed into inherent step changes and intervals of continuous change. Each continuous interval is approximated as a finite number of piecewise linear functions, which admits an analytical solution for perfectly mixed region. Using the Laplace transform, the temperature profiles in plug flow region with both the semi-infinite and adiabatic ends are successfully derived in terms of well-defined functions. The effect of end condition on the solution proves to be negligible under the practical operating conditions. For a Quadratic variation of inlet temperature, the approximate solution employing a moderate number of pieces agrees excellently with the exact solution.

• Speech Sciences
• /
• v.7 no.4
• /
• pp.11-26
• /
• 2000

In this paper, we propose the HMM with the MIN module. Because initial and re-estimated variance vectors are important elements for performance in HMM recognition systems, we propose a method which compensates for the mismatched statistical feature of training and test data. The MIN module function is a differentiable function similar to the sigmoid function. Unlike a continuous density function, it does not include variance vectors of the data set. The proposed hybrid HMM/MIN module is a unified network in which the observation probability in the HMM is replaced by the MIN module neural network. The parameters in the unified network are re-estimated by the gradient descent method for the Maximum Likelihood (ML) criterion. In estimating parameters, the variance vector is not estimated because there is no variance element in the MIN module function. The experiment was performed to compare the performance of the proposed HMM and the conventional HMM. The experiment measured an isolated number for speaker independent recognition.

• Bulletin of the Korean Mathematical Society
• /
• v.22 no.1
• /
• pp.17-22
• /
• 1985

B.M. Munshi and D.S. Bassan defined and developed the concept of super continuity in [5]. The concept has been investigated further by I. L. Reilly and M. K. Vamanamurthy in [6] where super continuity is characterized in terms of the semi-regularization topology. Super continuity is related to the concepts of .delta.-continuity and strong .theta.-continuity developed by T. Noiri in [7]. The purpose of this note is to derive relationships between super continuity and other strong continuity conditions and to develop additional properties of super continuous functions. Super continuity implies continuity, but the converse implication is false [5]. Super continuity is strictly between strong .theta.-continuity and .delta.-continuity and strictly between complete continuity and .delta.-continuity. The symbols X and Y will denote topological spaces with no separation axioms assumed unless explicity stated. The closure and interior of a subset U of a space X will be denoted by Cl(U) and Int(U) respectively and U is said to be regular open (resp. regular closed) if U=Int[Cl(U) (resp. U=Cl(Int(U)]. If necessary, a subscript will be added to denote the space in which the closure or interior is taken.

• Proceedings of the IEEK Conference
• /
• 2000.09a
• /
• pp.935-938
• /
• 2000

본 논문에서는 음성 인식을 이용하여 전화를 교환해주는 자동 교환 시스템을 구현하고, 성능을 평가하였다. 구현된 시스템에는 필수적인 음성인식 이외에도 DSP 진단 기능, 인식 대상 어휘의 추가 및 변경기능, 음성 수집 기능 등이 구현 되었다. SCHMM (Semi-Continuous Hidden Markov Model)을 이용한 전화망에서의 화자 독립 고립 단어 가변 어휘 인식을 대상으로 하였으며, 실시간 구현을 위하여 Texas Instrument 사의 TMS320C32를 사용하였다〔6〕. 인식 어휘는 부서명 및 인명이고 1300여 단어일 때, 인식 성능은 91.5%이다.

• Korean Journal of Mathematics
• /
• v.13 no.1
• /
• pp.83-90
• /
• 2005

A.H.Hamel proved the Ekeland's principle in a locally convex Hausdorff topological vector spaces by constructing the norm and applying the Ekeland's principle in Banach spaces. In this paper we show that the Ekeland's principle in a locally convex Hausdorff topological vector spaces can be proved directly by applying the famous general principle of H.Br$\acute{e}$zis and F.E.Browder.

• Journal of applied mathematics & informatics
• /
• v.28 no.1_2
• /
• pp.487-498
• /
• 2010

For some $\phi$-sequences $\phi_1$, $\phi_2$ and $\phi_3$, and $\kappa$-function $\kappa_1$, $\kappa_2$ and $\kappa_3$ with $\kappa_1^{-1}(x)\kappa_2^{-1}(x)\;{\geq}\;\kappa_3^{-1}(x)$ for $x\;{\geq}\;0$, the Luxemburg norm is lower semi-continuous on ${\kappa}{\phi}BV_0$, and some specialized equivalent conditions are considered.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 2007.10a
• /
• pp.305-306
• /
• 2007

• Journal of the Korean Mathematical Society
• /
• v.50 no.1
• /
• pp.95-109
• /
• 2013

The aim of this paper is to establish variational principle on cone metric spaces and to give some existence theorems of solutions for equilibrium problems on cone metric spaces. We give some equivalences of an existence theorem of solutions for equilibrium problems on cone metric spaces.

• The Journal of the Acoustical Society of Korea
• /
• v.11 no.6
• /
• pp.46-52
• /
• 1992

본 논문에서는 HMM 의 이산분포를 연속분포로 근사시키는 준 연속분포 HMM 에 의한 한국어 단어인식에 관하여 연구하였다. 이 모델의 생성과정에서는 입력벡터의 출력확률을 혼합 다차원 정규분 포로 가정하여 입력벡터의 확률함수와 코드위드의 심볼출력을 선형결합하므로써, 연속분포 모델로 근사 시켰으며, 단어인식과정에서는 생성모델에 의해 이산분포 모델에서 발생되는 양자와 왜곡을 감소시키므 로써 인식률을 향상시켰다. 이 방법을 평가하기 위하여 DDD 지역명을 대상으로 이산분포 HMM과 준연 속분포 HMM 의 비교실험을 수행하였다. 그 결과 준연속분포 HMM 에 의하여 이산분포 HMM 보다 향상된 인식률을 얻을 수 있었다.