• Journal of KSNVE
• /
• v.26 no.2
• /
• pp.10-17
• /
• 2016

• Korea Multimedia Society
• /
• v.12 no.1
• /
• pp.18-31
• /
• 2008

• Proceedings of the KWS Conference
• /
• 2000.04a
• /
• pp.117-120
• /
• 2000

• Transactions of Materials Processing
• /
• v.1 no.1
• /
• pp.5-10
• /
• 1992

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.9
• /
• pp.125-139
• /
• 2014

In this paper, I propose a response modeling with a Semi-Supervised Support Vector Regression (SS-SVR) algorithm. In order to increase the accuracy and profit of response modeling, unlabeled data in the customer dataset are used with the labeled data during training. The proposed SS-SVR algorithm is designed to be a batch learning to reduce the training complexity. The label distributions of unlabeled data are estimated in order to consider the uncertainty of labeling. Then, multiple training data are generated from the unlabeled data and their estimated label distributions with oversampling to construct the training dataset with the labeled data. Finally, a data selection algorithm, Expected Margin based Pattern Selection (EMPS), is employed to reduce the training complexity. The experimental results conducted on a real-world marketing dataset showed that the proposed response modeling method trained efficiently, and improved the accuracy and the expected profit.

• KIISE Transactions on Computing Practices
• /
• v.23 no.12
• /
• pp.698-702
• /
• 2017

Since the introduction of Deep Neural Networks to the Natural Language Processing field, two major approaches have been considered for modeling text. One method involved learning embeddings, i.e. the distributed representations containing abstract semantics of words or sentences, with the textual context. The other strategy consisted of composing the embeddings trained by the above to get embeddings of longer texts. However, most studies of the composition methods just adopt word embeddings without consideration of the optimal embedding unit and the optimal method of composition. In this paper, we conducted experiments to analyze the optimal embedding unit and the optimal composition method for modeling longer texts, such as documents. In addition, we suggest a new discourse-based composition to overcome the limitation of the sequential composition method on composing sentence embeddings.

• Explosives and Blasting
• /
• v.23 no.1
• /
• pp.47-54
• /
• 2005

An explosion modeling technique was developed by using the spherical discrete element code, $PFC^{3D}$, which can be used to model the dynamic stress wave propagation phenomenon. The modeling technique is simply based on an idea that the explosion pressure should be applied to a $PFC^{3D}$ particle assembly not in the form of an external force (body force), but in the form of a contact force (surface force). A test blast was conducted for a RC column, whose dimension was $600\times300\times1800$ in millimeters. The initial velocities of the surface movements were measured to be in the range of $14\~18\;m/s$ with the initiation times of $1.5\~2.0m$. Then the blasting procedure was simulated by using the modeling technique. The particle assembly representing the concrete was made of cement mortar and coarse aggregates, whose mirco-properties were obtained from the calibration processes. As a result, the modeling technique developed in this study made it possible for the burden to move with the velocity of $17\~24\;m/s$, which are slightly higher values compared to those of the test blast.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.3
• /
• pp.105-115
• /
• 2015

In semiconductor etching simulation, The source modeling for generating plasma species is required. In this paper, we modeled the source of plasma etching process with probability distribution and the feature profile with simple geometry objects, then got the flux on the feature profile. The distance between the source and the cell on the modeling parameters of the source, there are a number of particles to be emitted from a source, there is a number (area of the cell) of the cell on the profile with additional parameters to give the calculation of flux. The flux error ratio on both gaussian(Incident Flux) and cosine probability distribution(Incident Neutral Flux) is much decreased as the number of ray is increased but the processing time is more increased than that. The increase of the number of cell and distance makes increase the flux error ratio and the processing time moderately. In view of the processing time through the experimental results in this paper, it is possible to analogize the calculation of appropriate fluxes.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.43 no.1
• /
• pp.23-32
• /
• 2015

Modeling of fighter pilots, which is a fundamental technology for war games using defense M&S (Modeling & Simulation) becomes one of the prominent research issues as the importance of defense M&S increases. Especially, the recent accumulation of combat logs makes it possible to adopt statistical learning methods to pilot modeling, and an HMM (Hidden Markov Model) which is able to utilize the sequential characteristic of combat logs is suitable for the modeling. However, since an HMM works only by using one type of features, discrete or continuous, to apply an HMM to heterogeneous features, type integration is required. Therefore, we propose a dPCA-HMM method, where dPCA (Discrete Principal Component Analysis) is combined with an HMM for the type integration. From experiments conducted on combat logs acquired from a simulator furnished by agency for defense development, the performance of the proposed model is evaluated and was satisfactory.

• Journal of Aerospace System Engineering
• /
• v.11 no.5
• /
• pp.6-12
• /
• 2017

This paper deals with mathematical modeling of a 75-ton open-cycle Liquid Propellant Rocket Engine (LPRE) and the steady state simulation based on a nominal operating point. Each component of open-cycle LPRE may be classified into seven major categories using thermodynamics and dynamics characteristics. To simplify the simulation model of LPRE in this paper, we used four govern equations with assuming no heat transfer process. We confirmed the mathematical model of LPRE by using the error ratio and comparing the experiment data and simulation data in steady state, and checked the stability with the linearized model. Finally, we demonstrated the simulation model as compared to the transient response of experimental data.