• 한국음향학회지
• /
• 제32권3호
• /
• pp.227-236
• /
• 2013

본 논문에서는 SAOC의 Karaoke 모드의 출력 신호 내에 존재하는 잔여 보컬 성분을 추정하여 억제시킴으로써 음질을 향상시킬 수 있는 알고리듬을 제안하였다. 잔여 보컬 성분은 Karaoke 모드 환경으로 합성된 신호와 Solo 모드로 새로 합성된 신호를 서로 교차 예측하여 추정될 수 있다. 그러나, 두 신호는 모두 같은 다운 믹스 신호로부터 합성되는 신호이므로, 두 신호간의 높은 상관성으로 인하여 가라오케 신호내의 잔여 보컬 성분뿐만 아니라 음악 성분도 함께 제거된다. 이러한 열화를 해결하기 위해, 본 논문에서는 교차 예측 과정에서 심리 음향적 특성을 고려한 예측 방해 신호를 적용하였으며, 이 신호의 크기는 심리음향모델의 마스킹 특성에 따라 음악적 음질의 열화가 최소화되도록 적응적으로 설정되었다. 실험은 보컬 객체가 포함된 음악 신호에 대해서 객관적 및 주관적 음질평가를 수행하였으며, 전체적으로 성능 향상이 있음을 확인하였다.

• 한국분무공학회지
• /
• 제18권1호
• /
• pp.35-43
• /
• 2013

The empirical correlations for the prediction of breakup length of liquid jet in uniform cross flow are reviewed and classified in this study. The breakup length of liquid jets in cross flow was normally discussed in terms of the distances from the nozzle exit to the column breakup location in the x and y directions, called as column fracture distance and column fracture height, respectively. The empirical correlations for the prediction of column fracture distance can be classified as constant form, momentum flux ratio form, Weber number form and other parameter form, respectively. In addition, the empirical correlations for the prediction of column fracture height can be grouped as momentum flux ratio form, Weber number form and other parameter form, respectively. It can be summarized that the breakup length of liquid jet in a cross flow is a basically function of the liquid to air momentum flux ratio. However, Weber number, liquid-to-air viscosity ratio and density ratio, Reynolds number or Ohnesorge number were incorporated in the empirical correlations depending on the investigators. It is clear that there exist the remarkable discrepancies of predicted values by the existing correlations even though many correlations have the same functional form. The possible reasons for discrepancies can be summarized as the different experimental conditions including jet operating condition and nozzle geometry, measurement and image processing techniques introduced in the experiment, difficulties in defining the breakup location etc. The evaluation of the existing empirical correlations for the prediction of breakup length of liquid jet in a uniform cross flow is required.

• 대한전기학회논문지:전력기술부문A
• /
• 제55권11호
• /
• pp.485-491
• /
• 2006

In this paper, we study the problem of model selection for Support Vector Machine(SVM) predictor for short-term load forecasting. The model selection amounts to tuning SVM parameters, such as the cost coefficient C and kernel parameters and so on, in order to maximize the prediction performance of SVM. We propose that Cross-Validation method can be used as a model selection algorithm for SVM-based load forecasting technique. Through the various experiments on several data sets, we found that the difference between the prediction error of SVM using Cross-Validation and that of ideal SVM is less than 5%. This shows that SVM parameters for load forecasting can be efficiently tuned by using Cross-Validation.

• Ocean Systems Engineering
• /
• 제11권1호
• /
• pp.43-58
• /
• 2021

In the present work, a 3D numerical model is proposed to study local scouring around single vertical piers with different cross-section shapes under steady-current flow. The model solves the flow field and sediment transport processes using a coupled approach. The flow field is obtained by solving the Unsteady Reynolds Averaged Navier-Stokes (URANS) equations in combination with the k-ω SST turbulence closure model and the sediment transport is considered using both bedload and suspended load models. The proposed model is validated against the empirical measurements of local scour around single vertical piers with circular, square, and diamond cross-section shapes obtained from the literature. The measurement of scour depth in equilibrium condition for the simulations reveal the differences of 4.6%, 6.7% and 13.1% from the experimental measurements for the circular, square, and diamond pier cases, respectively. The model displayed a remarkable performance in the prediction of scour around circular and square piers where horseshoe vortices (HSVs) have a leading impact on scour progression. On the other hand, the maximum deviation was found in the case of the diamond pier where HSVs are weak and have minimum impact on the formation of local scour. Overall, the results confirm that the prediction capability of the present model is almost independent of the strength of the formed HSVs and pier cross-section shapes.

• Nuclear Engineering and Technology
• /
• 제54권4호
• /
• pp.1439-1448
• /
• 2022

Background: We investigated the feasibility of in vitro radiosensitivity prediction with gene expression using deep learning. Methods: A microarray gene expression of the National Cancer Institute-60 (NCI-60) panel was acquired from the Gene Expression Omnibus. The clonogenic surviving fractions at an absorbed dose of 2 Gy (SF2) from previous publications were used to measure in vitro radiosensitivity. The radiosensitivity prediction model was based on the convolutional neural network. The 6-fold cross-validation (CV) was applied to train and validate the model. Then, the leave-one-out cross-validation (LOOCV) was applied by using the large-errored samples as a validation set, to determine whether the error was from the high bias of the folded CV. The criteria for correct prediction were defined as an absolute error<0.01 or a relative error<10%. Results: Of the 174 triplicated samples of NCI-60, 171 samples were correctly predicted with the folded CV. Through an additional LOOCV, one more sample was correctly predicted, representing a prediction accuracy of 98.85% (172 out of 174 samples). The average relative error and absolute errors of 172 correctly predicted samples were 1.351±1.875% and 0.00596±0.00638, respectively. Conclusion: We demonstrated the feasibility of a deep learning-based in vitro radiosensitivity prediction using gene expression.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 1998년도 추계 학술대회논문집
• /
• pp.36-41
• /
• 1998

The cross-flow fan performance and its sound noise characteristics are predicted by computational methods. The unsteady incompressible Navier-Stokes equations in moving coordinates are solved by a SMAC method on unstructured triangular meshes, using a sliding mesh technique at the interface between the domain rotating with blades and the rest stationary part. The computationally predicted fan performance was favorably compared with experiment, and some numerical aspects of simulating the cross-flow fan are discussed. With the computed unsteady flow field, aeroacoustic sound noise of the fan is predicted by the Lighthill-Curie equation. The unsteady surface pressure fluctuations on stabilizer enables a prediction of BPF noise of the uniform pitch blade fan quite accurately. The aeroacoustic sound noise characteristics of both uniform and random pitch blade fans are also examined by SPL spectrum analysis.

• Communications for Statistical Applications and Methods
• /
• 제13권1호
• /
• pp.151-165
• /
• 2006

It is important to estimate the true misclassification rate of a given classifier when an independent set of test data is not available. Cross-validation and bootstrap are two possible approaches in this case. In related literature bootstrap estimators of the true misclassification rate were asserted to have better performance for small samples than cross-validation estimators. We compare the two estimators empirically when the classification rule is so adaptive to training data that its apparent misclassification rate is close to zero. We confirm that bootstrap estimators have better performance for small samples because of small variance, and we have found a new fact that their bias tends to be significant even for moderate to large samples, in which case cross-validation estimators have better performance with less computation.

• Communications for Statistical Applications and Methods
• /
• 제2권2호
• /
• pp.146-154
• /
• 1995

By using simple Taylor expansion, we derive an easy bias-correction rule for the apparent prodiction error of the predictor defined by the general M-estimators with respect to an arbitrary measure of prediction error. Our method has a considerable computational advantage over the previous methods based on the resampling thchnique such as Cross-validaton and Boothtrap. Connections with AIC, Cross-Validation and Boothtrap are discussed too.

• 대한의용생체공학회:의공학회지
• /
• 제33권1호
• /
• pp.1-7
• /
• 2012

Hemorrhagic shock is a common cause of death in emergency rooms. Early diagnosis of hemorrhagic shock makes it possible for physicians to treat patients successfully. Therefore, the purpose of this study was to select an optimal survival prediction model using physiological parameters for the two analyzed periods: two and five minutes before and after the bleeding end. We obtained heart rates, mean arterial pressures, respiration rates and temperatures from 45 rats. These physiological parameters were used for the training and testing data sets of survival prediction models using an artificial neural network (ANN) and support vector machine (SVM). We applied a 5-fold cross validation method to avoid over-fitting and to select the optimal survival prediction model. In conclusion, SVM model showed slightly better accuracy than ANN model for survival prediction during the entire analysis period.

• Communications for Statistical Applications and Methods
• /
• 제5권2호
• /
• pp.277-285
• /
• 1998

The smoothing parameter $\lambda$ in smoothing spline regression is usually selected by minimizing cross-validation (CV) or generalized cross-validation (GCV). But, simple CV or GCV is poor candidate for estimating prediction error. We defined MGCV (Multifold Generalized Cross-validation) as a criterion for selecting smoothing parameter in smoothing spline regression. This is a version of cross-validation using $leave-\kappa-out$ method. Some numerical results comparing MGCV and GCV are done.