• Journal of the Earthquake Engineering Society of Korea
• /
• v.11 no.2 s.54
• /
• pp.105-113
• /
• 2007

In the current research, an algorithm of performance-based seismic retrofit design of reinforced concrete columns using FRP jacket has been proposed. For exact prediction of the nonlinear flexural analysis or FRP composite RC members, multiaxial constitutive laws of concrete and composite materials have been presented. For seismic retrofit design, an algorithm of direct displacement-based design method (DDM) proposed by Chopra and Goel (2001) has been newly applied to determine the design thickness of FRP jacket in seismic retrofit of reinforced concrete columns. To compare with the displacement coefficient method (DCM), the DDM gives an accurate prediction of the target displacement in highly nonlinear region, since the DCM uses the elastic stiffness before reaching the yield load as the effective stiffness but the DDM uses the secant stiffness.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.11
• /
• pp.1265-1270
• /
• 2015

For the sensor node or collection node operating with a battery in a wireless sensor network, MAC protocols with improved energy efficiency are important performance factors. In this paper, in order to improve the restrictive capability in accordance with the fixed activity period of the duty cycle technology in the MAC protocol for wireless sensor networks, we propose a periodic adaptive wakeup technique based on receive prediction. The proposed technique is through a performance evaluation using the CC2500 RF transceiver and C8051F330 microcontroller based wireless node, to analyze the minimum active period. As a result, it was confirmed that it is possible to improve energy efficiency by adaptively changing the sleep period in accordance with the change of period.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.401-410
• /
• 2019

In this paper, we propose a reference picture generation method for Inter-layer prediction based deep learning to improve the SHVC coding performance. A description will be given of a structure for performing filtering using a VDSR network on a DCT-IF based upsampled picture to generate a new reference picture and a training method for generating a reference picture between SHVC Inter-layer. The proposed method is implemented based on SHM 12.0. In order to evaluate the performance, we compare the method of generating Inter-layer predictor by applying dictionary learning. As a result, the coding performance of the enhancement layer showed a bitrate reduction of up to 13.14% compared to the method using dictionary learning, a bitrate reduction of up to 15.39% compared to SHM, and a bitrate reduction of 6.46% on average.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.2
• /
• pp.207-213
• /
• 2024

In this paper, we study how to improve the performance of a machine learning-based automatic music transcription model by adding musical information to the input data. Where, the added musical information is information on the number of pitches that occur in each time frame, and which is obtained by counting the number of notes activated in the answer sheet. The obtained information on the number of pitches was used by concatenating it to the log mel-spectrogram, which is the input of the existing model. In this study, we use the automatic music transcription model included the four types of block predicting four types of musical information, we demonstrate that a simple method of adding pitch number information corresponding to the music information to be predicted by each block to the existing input was helpful in training the model. In order to evaluate the performance improvement proceed with an experiment using MIDI Aligned Piano Sounds (MAPS) data, as a result, when using all pitch number information, performance improvement was confirmed by 9.7 % in frame-based F1 score and 21.8 % in note-based F1 score including offset.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.48 no.5
• /
• pp.323-334
• /
• 2020

Since boundary layer transition has a significant impact on the aero-thermodynamic performance of hypersonic flight vehicles, capability of accurate prediction of transition location is essential for design and performance analysis. In this study, γ-Re_θt model is improved to predict transition of hypersonic boundary layers and validated. A coefficient in the production term of the intermittency transport equation that affects the transition onset location is constructed and applied as a function of Mach number, wall temperature, and freestream stagnation temperature based on the similarity numerical solution of compressible boundary layer. To take into account a Mach number dependency of transition onset momentum thickness Reynolds number and transition length, additional correlation equations are determined as function of Mach number and applied to Re_θc and F_length correlations of the baseline model. The suggested model is implemented to a commercial CFD code in consideration of practical use. Analysis of hypersonic flat plate and circular cone boundary layers is carried out by using the model for validation purpose. An improvement of prediction capability with respect to variation of Mach number and unit Reynolds number is identified from the comparison with experimental data.

• Journal of KIISE:Computer Systems and Theory
• /
• v.35 no.2
• /
• pp.75-83
• /
• 2008

Joye et al. introduced a new prime generation algorithm (JPV algorithm hereafter), by removing the trial division from the previous combined prime generation algorithm (combined algorithm hereafter) and claimed that JPV algorithm is $30{\sim}40%$ faster than the combined algorithm. However, they only compared the number of Fermat-test calls, instead of comparing the total running times of two algorithms. The reason why the total running times could not be compared is that there was no probabilistic analysis on the running time of the JPV algorithm even though there was a probabilistic analysis for the combined algorithm. In this paper, we present a probabilistic analysis on the running time of the JPV algorithm. With this analytic model, we compare the running times of the JPV algorithm and the combined algorithm. Our model predicts that JPV algorithm is slower than the combined algorithm when a 512-bit prime is generated on a Pentium 4 system. Although our prediction is contrary to the previous prediction from comparing Fermat-test calls, our prediction corresponds to the experimental results more exactly. In addition, we propose a method to improve the JPV algorithm. With this method, the JPV algorithm can be comparable to the combined algorithm with the same space requirement.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_3
• /
• pp.1031-1042
• /
• 2023

In this study, a deep learning model was developed to predict the yield of cabbage and radish, one of the five major supply and demand management vegetables, using satellite images of Landsat 8. To predict the yield of cabbage and radish in Gangwon-do from 2015 to 2020, satellite images from June to September, the growing period of cabbage and radish, were used. Normalized difference vegetation index, enhanced vegetation index, lead area index, and land surface temperature were employed in this study as input data for the yield model. Crop yields can be effectively predicted using satellite images because satellites collect continuous spatiotemporal data on the global environment. Based on the model developed previous study, a model designed for input data was proposed in this study. Using time series satellite images, convolutional neural network, a deep learning model, was used to predict crop yield. Landsat 8 provides images every 16 days, but it is difficult to acquire images especially in summer due to the influence of weather such as clouds. As a result, yield prediction was conducted by splitting June to July into one part and August to September into two. Yield prediction was performed using a machine learning approach and reference models , and modeling performance was compared. The model's performance and early predictability were assessed using year-by-year cross-validation and early prediction. The findings of this study could be applied as basic studies to predict the yield of field crops in Korea.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2006.06a
• /
• pp.176-182
• /
• 2006

본 연구에서는 기존 Bagging Predictors에 수정을 가한 Modified Bagging Predictors를 이용하여 SOHO 에 대한 부도예측 모델을 제시한다. 대기업 및 중소기업에 대한 기압부도예측 모델에 대한 많은 선행 연구가 있어왔지만 SOHO 만의 기업부도예측 모델에 관한 연구는 미비한 상태이다. 금융기관들의 대출심사시 대기업 및 중소기업과는 달리 SOHO에 대한 대출심사는 이직은 체계화되지 못한 채 신용정보점수 등의 단편적인 요소를 사용하고 있는 것에 현실이고 이에 따라 잘못된 대출로 안한 금융기관의 부실화를 초래할 위험성이 크다. 본 연구에서는 실제 국내은행의 SOHO 데이터 집합이 사용되었다. 먼저 기업부도 예측 모델에서 우수하다고 연구되어진 인공신경망과 의사결정나무 추론 기법을 적용하여 보았지만 만족할 만한 성과를 이쓸어내지 못하여, 기존 기업부도예측 모델연구에서 적용이 미비하였던 Bagging Predictors와 이를 개선한 Modified Bagging Predictors를 제시하고 이를 적용하여 보았다. 연구결과,; SOHO 부도예측에 있어서 본 연구에서 제시한 Modified Bagging Predictors 가 인공신경망과 Bagging Predictors등의 기존 기법에 비해서 성과가 향상됨을 알 수 있었다. 제시된 Modified Bagging Predictors의 유용성을 확인하기 위해서 추가적으로 대수의 공개 데이터 집합을 활용하여 성능을 비교한 결과 Modified Bagging Predictors 가 기존의 Bagging Predictors 에 비해 일관적으로 성과가 향상됨을 알 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.7-10
• /
• 2022

본 연구는 국내 주식의 intraday 가격변화를 딥러닝 모형들로 예측하고 그 예측모형을 이용한 매매전략 딥러닝 모형을 제안한다. 주식의 intraday 가격변화에 따라서, 고빈도 매매, 주문집행문제 (order execution problem), 자동화 매매 등과 같은 intraday 주식 트레이딩의 수익률이 달라지기 때문에, 주식의 intraday 가격변화 예측은 주식 투자에 있어서 중요하다. 해외 시장에 대해서는 인공지능 등을 이용한 intraday 가격변화 예측 연구가 활발히 이루어졌지만, 국내의 경우 관련한 연구가 드물어 그 효용성이 명확히 드러나지 않았었다. 그에 따라서, KOSPI 50의 구성 종목에 대하여 정준의(canonical) 딥러닝 모형들을 적용하여 예측 성능을 비교한다. 또한, 그 예측모형들을 활용하여 간소화된 주문집행문제에서의 매매전략 딥러닝 모형을 제안한다. 그리고, 제안한 매매전략 딥러닝 모형을 KOSPI 50의 구성 종목에 대하여 실험하여, 제안한 방법론이 유효함을 밝힌다. 제시된 모형을 실제 주식 매매에 직접 적용하여 수익성을 개선을 기대할 수 있고, 사람이 직접 거래할지라도 효과적인 보조 지표가 될 수 있기에 본 논문은 실용적 의미를 지닌다.

• Journal of the Korean earth science society
• /
• v.45 no.1
• /
• pp.19-36
• /
• 2024

The global seasonal forecasting system version 6 (GloSea6) operated by the Korea Meteorological Administration for 1- and 3-month prediction products does not include complex atmospheric chemistry-aerosol physical processes (UKCA). In this study, low-resolution GloSea6 and GloSea6 coupled with UKCA (GloSea6-UKCA) were installed in a CentOS-based Linux cluster system, and preliminary prediction results for the spring of 2000 were examined. Low-resolution versions of GloSea6 and GloSea6-UKCA are highly needed to examine the effects of atmospheric chemistry-aerosol owing to the huge computational demand of the current high resolution GloSea6. The spatial distributions of the surface temperature and daily precipitation for April 2000 (obtained from the two model runs for the next 75 days, starting from March 1, 2000, 00Z) were compared with the ERA5 reanalysis data. The GloSea6-UKCA results were more similar to the ERA5 reanalysis data than the GloSea6 results. The surface air temperature and daily precipitation prediction results of GloSea6-UKCA for spring, particularly over East Asia, were improved by the inclusion of UKCA. Furthermore, compared with GloSea6, GloSea6-UKCA simulated improved temporal variations in the temperature and precipitation intensity during the model integration period that were more similar to the reanalysis data. This indicates that the coupling of atmospheric chemistry-aerosol processes in GloSea6 is crucial for improving the spring predictions over East Asia.