• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06e
• /
• pp.399-402
• /
• 1998

본 논문에서는 환자의 음성을 정상, 양성종양, 악성종양으로 분류하는 실험을 켑스트럼 파라미터를 통한 음원분리와 신경회로망을 이용하여 수행하고 그 결과를 보고한다. 기존의 장애음성 데이터베이스에는 정상음성과 양성종양의 경우만 수록되어 있었고 외국의 환자들을 대상으로 한 경우만 있었기 때문에 국내의 환자들에게 직접 적용할 경우 어떠한 결과가 나올지 예측하기가 어려웠다. 최근 부산대학교 이비인후과팀에서 수집한 국내의 정상, 양성, 악성종양의 경우에 대한 데이터베이스를 분석하고 신경회로망에 의해 분류함으로써 사람의 음성신호만에 의한 후두질환이 식별이 가능하였다. 본 실험에서는 식별 파라미터로 음성신호의 선형예측오차신호에 관한 켑스트럼으로부터 음원비인 HNRR을 구하여 Jitter, Shimmer와 함께 사용하였다. 신경회로망은 입, 출력 층과 한 개의 은닉층을 갖는 다층신경망을 이용하였으며, 식별은 두단계로 나누어 정상과 비정상을 분류한 후 다시 비정상을 양성과 악성으로 분류하였다[1].

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2003.09a
• /
• pp.447-453
• /
• 2003

연구지역은 경남 양산시 상북면 삼삼리 일원으로 양산단층과 인접하며 급한 경사의 자연사면으로 이루어져 있다. 수치지형자료를 이용하여 고도값을 갖는 자료층을 추출한 후 10m$\times$10m 크기의 격자별로 DEM으로 변환하고, 이로부터 단위 격자의 경사도, 경사방향도, 음영도를 작성하고 흐름분배 알고리즘을 이용하여 설정된 격자별로 상부사면의 기여면적과 습윤지수를 산정하였다. 높은 습윤지수를 보이는 지역은 북서-남동 방향의 선형구조로 나타나며, 국소지역의 전단단열 특성과 일치한다. 한편 연구지역에서 전기비저항 탐사를 실시하여 높은 습윤지수를 나타내는 지역과 저비저항 이상대의 분포를 비교한 결과 습윤지수가 높은 지역은 모두 낮은 전기비저항 이상대로 관찰된다. 즉 습윤지수는 암석의 화학적 풍화를 수반하는 파쇄대의 분포와 일치하며 이를 토대로 국소지역에서 파쇄대의 예측과 암석의 안정성을 예측하는 방법으로 이용될 수 있다.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.10 no.4
• /
• pp.165-173
• /
• 1998

Analytical solutions to predict the movement rate of submerged mound are derived using the convection coefficient and the joint distribution function of wave heights and periods. Assuming that the sediment is moved onshore due to the velocity asymmetry of Stokes' second order nonlinear wave theory, the micro-scale bedload transport equation is applied to the sediment conservation. The nonlinear convection-diffusion equation can then be obtained which governs the migration of submerged mound. The movement rate decreases exponentially with increasing the water depth, but the movement rate tends to increase as the spectral width parameter, $ u$ increases. In comparison of the analytical solution with the measured data, it is found that the analytical solution overestimates the movement rate. However, the agreement between the analytical solution and the measured data is encouraging since this over-estimation may be due to the inaccuracy of input data and the limitation of sediment transport model. In particular, the movement rates with respect to the water depth predicted by the analytical solution are in very good agreement with the estimated result using the discritization technique with the hindcast wave data.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.22 no.6
• /
• pp.597-605
• /
• 2009

The present study has been proposed a model for the in-plane shear behavior of reinforced(Engineered Cementitious Composite(ECC) panels under biaxial stress states. The model newly considers the high-ductile tensile characteristic of cracked ECC by its multiple micro-cracking mechanism, the compressive strain-softening characteristic of cracked ECC, and the shear transfer mechanism in the cracked interface of ECC element. A series of numerical analyses were performed, and the predicted curves were compared with experimental results. The proposed in-plane shear model, R-ECC-MCFT, was found to be well matched with the experimental results, and it was also demonstrated that reinforced ECC panel showed more improved in-plane shear strength and post peak behavior, in comparing with the conventional reinforced concrete panel.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.15 no.1
• /
• pp.51-58
• /
• 2003

The magnitude of evanescent modes in terms of dynamics it investigated in case that the transformation of water waves is predicted by the linear wave theory. For the waves propagating over two steps, the eigenfunction expansion method is used to predict the amplitudes of reflected and transmitted waves by the component of evanescent modes as well as propagating modes. Then. the relative importance of evanescent modes to the propagating modes is investigated. The numerical experiments find that the evanescent modes are pronounced at the relative water depth of k$_1$h$_1$＝0.11$\pi$ and the water depth ratio of h$_2$/h$_1$ close to zero.

• Journal of the Korean GEO-environmental Society
• /
• v.7 no.1
• /
• pp.41-53
• /
• 2006

In this study, estimation of nonlinear adsorption isotherms(Langmuir & Freundlich adsorption isotherm) and advection-dispersion model parameters was conducted using genetic algorithm(GA) for Zn and Cd adsorption. Estimated parameters of nonlinear adsorption isotherms, which were obtained from the optimization process using genetic algorithm(GA), are nearly same with the parameters obtained from a linearization process of the nonlinear isotherms. Estimated effective diffusion coefficients, which were obtained from a finite element analysis of the advection-dispersion model and an optimization procedure using the genetic algorithm, for the metals were approximately in the order of $10^{-7}cm^2/s$ which could be obtained based on the linear distribution coefficient. The effective diffusion coefficients based on the nonlinear retardation factors were in the range of $10^{-6}{\sim}10^{-5}cm^2/s$. As a result, the correlation coefficient obtained between the measured and calculated concentration was over 0.9 which means that the genetic algorithm should be successfully applied to estimate the unknown parameters of the nonlinear adsorption isotherms and advection-dispersion model.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.16 no.4
• /
• pp.1-8
• /
• 2012

Domestic steel moment resisting frames were designed in accordance with the former KBC2005 and the current KBC2009, and then their seismic performance was evaluated in accordance with FEMA355F by utilizing nonlinear dynamic analysis. The results from the procedure in FEMA355F were different with those from the capacity spectrum method utilizing nonlinear static push-over analysis. In particular, the domestic steel moment resisting frames have a weak panel zone, so their behavior can be estimated more precisely by nonlinear dynamic analysis. The domestic steel moment resisting frames satisfied the performance goal if located at a site class $S_B$ or $S_C$, regardless of the story number and the response modification factor. However, if they are located at a site class $S_D$ or $S_E$, performance goal satisfaction cannot be guaranteed. No matter what standard is used for the design, KBC2005 or KBC2009, the domestic steel moment resisting frames may possess satisfactory seismic performance if the site condition is relatively good.

• Journal of the Korean Society for Advanced Composite Structures
• /
• v.2 no.2
• /
• pp.1-6
• /
• 2011

This study aims at predicting structural behaviors of RC (Reinforced Concrete) beams retrofitted with hybrid FRPs (Fiber Reinforced Polymers). Toward this goal, structural analysis for the RC beams retrofitted with hybrid FRPs are performed and validated using existing experimental data. For the analysis, failure models due to debonding of FRPs and concrete separation are implemented within FE (Finite Element) model, based on Smith and Teng, model, and Teng and Yao model, respectively. Nonlinear material and geometrical effects are also included in the analysis. The suggested modeling approaches are able to predict structural behaviors of RC beams retrofitted with hybrid FRPs similar to the experimental data, however, a numerical model needs to be developed in order to predict failure strength of RC beams retrofitted with hybrid FRPs accurately.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.19 no.8
• /
• pp.77-84
• /
• 2005

The considerations for reminder life of transmission line is gradually higher. It is requisite for investigation of ACSR life to test tensile load of ACSR as a fundamental data. It is vary important to analysis correlations between results of tensile load testing and elapsed years. Estimation of ACSR life can be obtained by statistics processing using mechanical experimental results. It is a general method to use regression analysis as a statistics processing technique. In this paper, we did experiment on tensile strength of ACSR by using a new and due to flame for artificial fire, and gathering due to a flame. The limit of life estimation is decided by basic line using twenty percentage reduction of rate tensile strength. This basic line is like to results of Canada Ontario Hydro-research. There are $480[mm^2]$ ACSR which are experimented on this study.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.3
• /
• pp.176-182
• /
• 2024

In this study, we compared and analyzed various methods of missing data handling to build a machine learning model that can effectively analyze and predict time series data with a high percentage of missing values. For this purpose, Predictive State Model Filtering (PSMF), MissForest, and Imputation By Feature Importance (IBFI) methods were applied, and their prediction performance was evaluated using LightGBM, XGBoost, and Explainable Boosting Machines (EBM) machine learning models. The results of the study showed that MissForest and IBFI performed the best among the methods for handling missing values, reflecting the nonlinear data patterns, and that XGBoost and EBM models performed better than LightGBM. This study emphasizes the importance of combining nonlinear imputation methods and machine learning models in the analysis and prediction of time series data with a high percentage of missing values, and provides a practical methodology.