• 한국컴퓨터정보학회논문지
• /
• 제21권5호
• /
• pp.79-89
• /
• 2016

In this paper, we proposed the object classification method using genetic and dynamic random forest consisting of optimal combination of unit tree. The random forest can ensure good generalization performance in combination of large amount of trees by assigning the randomization to the training samples and feature selection, etc. allocated to the decision tree as an ensemble classification model which combines with the unit decision tree based on the bagging. However, the random forest is composed of unit trees randomly, so it can show the excellent classification performance only when the sufficient amounts of trees are combined. There is no quantitative measurement method for the number of trees, and there is no choice but to repeat random tree structure continuously. The proposed algorithm is composed of random forest with a combination of optimal tree while maintaining the generalization performance of random forest. To achieve this, the problem of improving the classification performance was assigned to the optimization problem which found the optimal tree combination. For this end, the genetic algorithm methodology was applied. As a result of experiment, we had found out that the proposed algorithm could improve about 3~5% of classification performance in specific cases like common database and self infrared database compare with the existing random forest. In addition, we had shown that the optimal tree combination was decided at 55~60% level from the maximum trees.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제10권2호
• /
• pp.57-64
• /
• 2021

자율주행차 시대가 도래하면서 ECU (Electronic Control Unit)는 점차 고도화되고 있고, 이에 따라 차량에서 정확한 데이터를 추출하고 분석하려는 연구가 다양하게 시도되어 왔다. 그러나 ECU는 차량 제조사별로 상이한 프로토콜을 가지고 있어 상용 단말기로는 정확한 데이터 추출과 분석이 어렵다. 본 연구에서는 정확한 차량 데이터를 추출하기 위하여 전용 펌웨어를 개발하여 차량의 2019년 1월부터 2월의 실제 주행데이터 53,580건의 데이터를 추출하였으며, 20회가 넘는 실제 도로 주행을 통해서 데이터의 정확도를 검증하였다. 이러한 데이터를 바탕으로 실시간 연료 소비량 예측 모델의 정확도를 높이기 위하여 스태킹 앙상블 기법을 이용하였다. 본 연구에서는 베이스 모델로 Ridge, Lasso, XGBoost, LightGBM이 사용되고 메타 모델은 Ridge가 사용되었으며, 예측 성능은 MAE 0.011, RMSE 0.017로 최적의 결과를 보였다.

• Bulletin of the Korean Chemical Society
• /
• 제32권11호
• /
• pp.3909-3913
• /
• 2011

In this paper we have carried out molecular dynamics simulations (MD) for model systems of liquid n-alkane oligomers ($C_{12}{\sim}C_{80}$) at high temperatures (~2300 K) in a canonical ensemble to calculate viscosity ${\eta}$, self-diffusion constants D, and monomeric friction constant ${\zeta}$. We found that the long chains of these n-alkanes at high temperatures show an abnormality in density and in monomeric friction constant. The behavior of both activation energies, $E_{\eta}$ and $E_D$, and the mass and temperature dependence of ${\eta}$, D, and ${\zeta}$ are discussed.

• Bulletin of the Korean Chemical Society
• /
• 제35권5호
• /
• pp.1469-1478
• /
• 2014

Use of acetonitrile in electrolytes promotes better operation of supercapacitors. Recent efforts show that electrolytes containing acetonitrile can also function in a wide range of operating temperatures. Therefore, this paper addresses the dielectric relaxation processes, structure and dynamic properties of the bulk acetonitrile at various temperatures. Systems of acetonitrile were modeled using canonical ensemble and simulated by employing Molecular Dynamics method. Results show that interactions among the molecules were correlated within a cut-off radius while parallel and anti-parallel arrangements are observed beyond this radius at relatively high and low temperatures respectively. Furthermore, effects of C-C-N and C-H bending modes were greatly appreciated on the power spectral density of time rate change of dipole-dipole correlations whereas frequency shifts were observed on all modes at the lowest temperature under consideration. Linear variations with temperature were depicted for reorientation times and self-diffusion coefficients. Shear viscosity was also computed with a good accuracy within a certain range of the temperature as well.

• 한국의류산업학회지
• /
• 제17권4호
• /
• pp.635-645
• /
• 2015

This study analyzed the range of motion of upper body in different configurations of firefighters' protective clothing and equipment. The purpose of this study was to understand the influence of firefighters' protective clothing and equipment over upper body motion in order to improve design of firefighters' protective clothing and equipment. 12 firefighters' upper body range of motion was analyzed while performing standing and walking trials in five different garment configurations including turnout ensemble, fire boots and the self-contained breathing apparatus. Analysis of upper body range of motion included spinal joints of L5S1, L4L3, T1C7, and C1Head. During standing trials, garment configurations caused a significant difference in range of motions at joints of L5S1, L4L3, T1C7, and C1Head. Analysis on the mean of range of motions at L5S1 and L4L3, showed that firefighters' waist bent forward significantly to a greater extent while they wore a self-contained breathing apparatus. A significantly increased range of motion was found for T1C7 and C1Head while carrying a self-contained breathing apparatus, which indicated an increase in the extension of the trunk and neck backward to stand upright and look squarely. A significant difference in range of motion was also found for L5S1 and L4L3 during walking trials.

• Smart Structures and Systems
• /
• 제26권2호
• /
• pp.241-251
• /
• 2020

Due to the undeniable importance of approximating the concrete compressive strength (CSC) in civil engineering, this paper focuses on presenting four novel optimizations of multi-layer perceptron (MLP) neural network, namely artificial bee colony (ABC-MLP), grasshopper optimization algorithm (GOA-MLP), shuffled frog leaping algorithm (SFLA-MLP), and salp swarm algorithm (SSA-MLP) for predicting this crucial parameter. The used dataset consists of 103 rows of information concerning seven influential parameters (cement, slag, water, fly ash, superplasticizer, fine aggregate, and coarse aggregate). In this work, the best-fitted complexity of each ensemble is determined by a population-based sensitivity analysis. The GOA distinguished its self by the least complexity (population size = 50) and emerged as the second time-effective optimizer. Referring to the prediction results, all tested algorithms are able to construct reliable networks. However, the SSA (Correlation = 0.9652 and Error = 1.3939) and GOA (Correlation = 0.9629 and Error = 1.3922) performed more accurately than ABC (Correlation = 0.7060 and Error = 4.0161) and SFLA (Correlation = 0.8890 and Error = 2.5480). Therefore, the SSA-MLP and GOA-MLP can be promising alternatives to laboratorial and traditional CSC evaluative methods.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2011년도 제40회 동계학술대회 초록집
• /
• pp.38-38
• /
• 2011

Covalently bonded halogen ligands possess unusual charge distributions, attracting both electrophilic and nucleophilic molecular ligands to form halogen bonds. In many biochemical systems, halogen bonds and hydrogen bonds coexist. The interplay between halogen and hydrogen bonds has been actively studied in various three-dimensional bulk molecular co-crystals. It was found that halogen bonds could be complementary to hydrogen bonds due to their similar bond strength and dissimilar directionality. In those ensemble-averaging approaches, however, it was not possible to extract local information such as individual bond configurations and nano-level domain structures, which is a crucial part of supramolecular studies. In this study, we directly visualize the individual molecular configuration of a brominated molecule and the role of halogen bonds on Au(111) using scanning tunneling microscopy. The precise arrangement of observed molecular structures was reproduced by first-principle studies and explained in the context of halogen and hydrogen bonds. We discuss the distances and the strengths of the observed halogen bonds and hydrogen bonds, which are consistent with previous bulk data.

• Journal of Information Processing Systems
• /
• 제13권5호
• /
• pp.1135-1148
• /
• 2017

Speech recognition is one of the fascinating fields in the area of Computer science. Accuracy of speech recognition system may reduce due to the presence of noise present in speech signal. Therefore noise removal is an essential step in Automatic Speech Recognition (ASR) system and this paper proposes a new technique called combined thresholding for noise removal. Feature extraction is process of converting acoustic signal into most valuable set of parameters. This paper also concentrates on improving Mel Frequency Cepstral Coefficients (MFCC) features by introducing Discrete Wavelet Packet Transform (DWPT) in the place of Discrete Fourier Transformation (DFT) block to provide an efficient signal analysis. The feature vector is varied in size, for choosing the correct length of feature vector Self Organizing Map (SOM) is used. As a single classifier does not provide enough accuracy, so this research proposes an Ensemble Support Vector Machine (ESVM) classifier where the fixed length feature vector from SOM is given as input, termed as ESVM_SOM. The experimental results showed that the proposed methods provide better results than the existing methods.

• Asia pacific journal of information systems
• /
• 제29권4호
• /
• pp.789-816
• /
• 2019

Text classification is a challenging task, especially when dealing with a huge amount of text data. The performance of a classification model can be varied depending on what type of words contained in the document corpus and what type of features generated for classification. Aside from proposing a new modified version of the existing algorithm or creating a new algorithm, we attempt to modify the use of data. The classifier performance is usually affected by the quality of learning data as the classifier is built based on these training data. We assume that the data from different domains might have different characteristics of noise, which can be utilized in the process of learning the classifier. Therefore, we attempt to enhance the robustness of the classifier by injecting the heterogeneous data artificially into the learning process in order to improve the classification accuracy. Semi-supervised approach was applied for utilizing the heterogeneous data in the process of learning the document classifier. However, the performance of document classifier might be degraded by the unlabeled data. Therefore, we further proposed an algorithm to extract only the documents that contribute to the accuracy improvement of the classifier.

• International Journal of Stem Cells
• /
• 제17권1호
• /
• pp.30-37
• /
• 2024

The lung is a complex organ comprising a branched airway that connects the large airway and millions of terminal gas-exchange units. Traditional pulmonary biomedical research by using cell line model system have limitations such as lack of cellular heterogeneity, animal models also have limitations including ethical concern, race-to-race variations, and physiological differences found in vivo. Organoids and on-a-chip models offer viable solutions for these issues. Organoids are three-dimensional, self-organized construct composed of numerous cells derived from stem cells cultured with growth factors required for the maintenance of stem cells. On-a-chip models are biomimetic microsystems which are able to customize to use microfluidic systems to simulate blood flow in blood channels or vacuum to simulate human breathing. This review summarizes the key components and previous biomedical studies conducted on lung organoids and lung-on-a-chip models, and introduces potential future applications. Considering the importance and benefits of these model systems, we believe that the system will offer better platform to biomedical researchers on pulmonary diseases, such as emerging viral infection, progressive fibrotic pulmonary diseases, or primary or metastatic lung cancer.