• International journal of advanced smart convergence
• /
• 제13권2호
• /
• pp.80-87
• /
• 2024

This paper focuses on improving accuracy in constrained computing settings by employing the ReLU (Rectified Linear Unit) activation function. The research conducted involves modifying parameters of the ReLU function and comparing performance in terms of accuracy and computational time. This paper specifically focuses on optimizing ReLU in the context of a Multilayer Perceptron (MLP) by determining the ideal values for features such as the dimensions of the linear layers and the learning rate (Ir). In order to optimize performance, the paper experiments with adjusting parameters like the size dimensions of linear layers and Ir values to induce the best performance outcomes. The experimental results show that using ReLU alone yielded the highest accuracy of 96.7% when the dimension sizes were 30 - 10 and the Ir value was 1. When combining ReLU with the Adam optimizer, the optimal model configuration had dimension sizes of 60 - 40 - 10, and an Ir value of 0.001, which resulted in the highest accuracy of 97.07%.

• International Journal of Internet, Broadcasting and Communication
• /
• 제12권3호
• /
• pp.125-130
• /
• 2020

At the end of 1997, the volatility of the exchange rate intensified as the nation's exchange rate system was converted into a free-floating exchange rate system. As a result, managing the exchange rate is becoming a very important task, and the need for forecasting the exchange rate is growing. The exchange rate prediction model using the existing exchange rate prediction method, statistical technique, cannot find a nonlinear pattern of the time series variable, and it is difficult to analyze the time series with the variability cluster phenomenon. And as the number of variables to be analyzed increases, the number of parameters to be estimated increases, and it is not easy to interpret the meaning of the estimated coefficients. Accordingly, the exchange rate prediction model using artificial neural network, rather than statistical technique, is presented. Using DNN, which is the basis of deep learning among artificial neural networks, and LSTM, a recurrent neural network model, the number of hidden layers, neurons, and activation function changes of each model found the optimal exchange rate prediction model. The study found that although there were model differences, LSTM models performed better than DNN models and performed best when the activation function was Tanh.

• 자원리싸이클링
• /
• 제12권3호
• /
• pp.13-24
• /
• 2003

폐호도껍질을 원료로 활성탄을 제조하는 과정에서 활성화 온도, 활성화 시간, 활성화제의 양, 그리고 활성화제의 종류 등을 변수로 하여 활성화 특성을 조사하였다. 인산을 활성화제로 사용하여 제조된 활성탄은 그 흡착능이 온도가 증가함에 따라 증가하여 약 $550^{\circ}C$부근에서 최대 흡착능을 보였으며 그 수율은 온도 상승에 따라 지속적으로 감소하였다. 활성화 시간은 약 2시간 정도에서 최적의 조건을 보였으며 시간이 증가함에 따라 활성탄의 수율은 계속 감소하였다. 활성화제의 농도 증가에 따라 수율은 지속적으로 상승하였으며 흡착능 또한 증대되다가 약 1.5M $H_3PO_4$ 이상의 조건에서는 오히려 흡착능이 감소하였다. SEM으로 관찰한 조건에 따른 활성탄의 미세구조의 변화는 조건별 흡착능의 변화와 잘 일치되었으며 활성화제의 종류는 활성화 과정에서 중요한 영향을 미치는 것으로 조사되었다. 제조된 활성탄의 흡착특성을 파악하기 위해 $Cu^{2+}$ 이온을 흡착질로 하여 흡착반응을 조사한 결과, 흡착반응은 전체적으로 2차식을 따르는 것으로 관찰되었으며 흡착질의 초기 농도가 감소함에 따라 반응상수는 점차 증가하였다. 평형흡착량은 Freundlich Model 을 잘 따르는 것으로 나타났으며 온도별 흡착반응을 검토한 결과, 중금속 이온의 흡착은 흡열반응의 특성을 나타내었다. 흡착에 따른 Activation Energy는 약 13.07kcal/mol로 산출되었으며 van't Hoff Equation을 이용하여 흡착반응의 열역학적 인자들을 계산하였다.

• 동의생리병리학회지
• /
• 제24권5호
• /
• pp.872-876
• /
• 2010

The present study found that EA pre-treatment effectively attenuated both spinal ERK activation and hyperalgesia against carrageenan-induced inflammation, whereas EA co-treatment with carrageenan injection unexpectedly elevated ERK activation in a synergistic manner and virtually had no analgesic effect. Therefore, we have concluded that the molecular mechanism of EA anagesia may be related to the inhibition of spinal ERK activation. Further experiments are required to find the intermediate candidates which transmits the pain-alleviating signals on the way of inhibiting ERK activation by EA.

• Bulletin of the Korean Chemical Society
• /
• 제29권10호
• /
• pp.1920-1926
• /
• 2008

We investigated the C-H bond activation mechanism of aldimine by the [RhCl$(PPH_3)_3$] model catalyst using DFT B3LYP//SBKJC/6-31G*/6-31G on GAMESS. Due to their potential utility in organic synthesis, C-H bond activation is one of the most active research fields in organic and organometallic chemistry. C-H bond activation by a transition metal catalyst can be classified into two types of mechanisms: direct C-H bond cleavage by the metal catalyst or a multi-step mechanism via a tetrahedral transition state. There are three structural isomers of [RhCl$(PH_3)_2$] coordinated aldimine that differ in the position of chloride with respect to the molecular plane. By comparing activation energies of the overall reaction pathways that the three isomeric structures follow in each mechanism, we found that the C-H bond activation of aldimine by the [RhCl$(PH_3)_3$] catalyst occurs through the tetrahedral intermediate.

• 디지털융복합연구
• /
• 제9권1호
• /
• pp.107-114
• /
• 2011

In order to have the highest level of a certain society's technology be evaluated as digital contents technology the value concept of such technology's social utilization must be established while active investment on the technology takes place and makes it the subject of social capitalization. This study wishes to discuss the strategies and methods of establishing and managing IP holding company which requires business activation with digital contents technology at its base, research ways of vitalizing IP holding company to expand social utility values of contents technology, suggest necessary systemic improvements and investment activation methods, management structure, and governance structure by investigating ways to stimulate the industrialization of contents technology through the establishment and management of this IP holding company, and finally come up with a realistic measure to establish and manage a IP holding company. Strategies on commercialization of digital contents technology and acceleration of technology development, as well as activation of venture business set-ups will be analyzed and suggested based on such suggestions while IP holding company's digital contents technology investment activation model will be established to produce means to realize discovering superior contents companies and activation of investment, and activating high quality contents production for the global market.

• 한국전기전자재료학회논문지
• /
• 제29권1호
• /
• pp.17-22
• /
• 2016

The activation energy of a material is an important factor that significantly affects the lifetime and can be used to develop a degradation model. In this study, a thermal analysis was carried out to evaluate and collect quantitative data on the degradation of insulation materials like EPR and CSP used for nuclear power plant cables. The activation energy was determined from the relationship between log ${\beta}$ and 1/T based on the Flynn.Wall.Ozawa method, by a TGA test. The activation energy was also derived from the relationship between ln(t) and 1/T based on isothermal analysis, by an OIT test. The activation energy of EPR derived from thermal analysis was used to calculate the accelerated aging time corresponding to the number of years of use, employing the Arrhenius equation, and determine the elongation corresponding to the accelerated aging time.

• Biomolecules & Therapeutics
• /
• 제30권1호
• /
• pp.55-63
• /
• 2022

Oleanolic acid (OA), a natural pentacyclic triterpenoid, has been reported to exert protective effects against several neurological diseases through its anti-oxidative and anti-inflammatory activities. The goal of the present study was to evaluate the therapeutic potential of OA against acute and chronic brain injuries after ischemic stroke using a mouse model of transient middle cerebral artery occlusion (tMCAO, MCAO/reperfusion). OA administration immediately after reperfusion significantly attenuated acute brain injuries including brain infarction, functional neurological deficits, and neuronal apoptosis. Moreover, delayed administration of OA (at 3 h after reperfusion) attenuated brain infarction and improved functional neurological deficits during the acute phase. Such neuroprotective effects were associated with attenuation of microglial activation and lipid peroxidation in the injured brain after the tMCAO challenge. OA also attenuated NLRP3 inflammasome activation in activated microglia during the acute phase. In addition, daily administration of OA for 7 days starting from either immediately after reperfusion or 1 day after reperfusion significantly improved functional neurological deficits and attenuated brain tissue loss up to 21 days after the tMCAO challenge; these findings supported therapeutic effects of OA against ischemic stroke-induced chronic brain injury. Together, these findings showed that OA exerted neuroprotective effects against both acute and chronic brain injuries after tMCAO challenge, suggesting that OA is a potential therapeutic agent to treat ischemic stroke.

• 한국원자력학회:학술대회논문집
• /
• 한국원자력학회 1996년도 춘계학술발표회논문집(2)
• /
• pp.637-642
• /
• 1996

This paper presents the analysis results for the core degradation processes and the fission product release of the PHEBUS FPT0 experiment using MELCOR1.8.3. The objective of this study is to assess models associated with the core damage and fission product behavior in MELCOR. The calculation results were much improved through sensitivity studies. Thermal/hydraulic behavior in the core and the circuit was well predicted under the intact core geometry. In non-eutectic model case. the UO$_2$ dissolution model in the MELCOR always showed such a tendency that the resulting dissolved UO$_2$ mass was small at the highly oxidized condition due to the model logic. Total H$_2$ generation mass was underpredicted because the stiffner was not modeled and the liner in the shroud was not allowed to be oxidized in MELCOR. Some difficulties were found in modeling the activation product were solved by manipulating the RN input associated with the initial fission product inventory. These problem were occurred because there are no control rod model in MELCOR. Generally the fission product release ratio showed a similar trend compared with the measured data except the activation product. which have no model to simulate in MELCOR.

• 대기
• /
• 제31권1호
• /
• pp.73-83
• /
• 2021

Numerical weather prediction (NWP) models play an essential role in predicting weather factors, but using them is challenging due to various factors. To overcome the difficulties of NWP models, deep learning models have been deployed in weather forecasting by several recent studies. This study adapts long short-term memory (LSTM), which demonstrates remarkable performance in time-series prediction. The combination of LSTM model input of meteorological features and activation functions have a significant impact on the performance therefore, the results from 5 combinations of input features and 4 activation functions are analyzed in 9 Automated Surface Observing System (ASOS) stations corresponding to cities/islands/mountains. The optimized LSTM model produces better performance within eight forecast hours than Local Data Assimilation and Prediction System (LDAPS) operated by Korean meteorological administration. Therefore, this study illustrates that this LSTM model can be usefully applied to very short-term weather forecasting, and further studies about CNN-LSTM model with 2-D spatial convolution neural network (CNN) coupled in LSTM are required for improvement.