• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1290-1295
• /
• 2017

Document classification, a task of classifying the category of each document based on text, is one of the fundamental areas for natural language processing. Document classification may be used in various fields such as topic classification and sentiment classification. Neural network models for document classification can be divided into two categories: word-level models and character-level models that treat words and characters as basic units respectively. In this study, we propose a neural network model that combines character-level and word-level models to improve performance of document classification. The proposed model extracts the feature vector of each word by combining information obtained from a word embedding matrix and information encoded by a character-level neural network. Based on feature vectors of words, the model classifies documents with a hierarchical structure wherein recurrent neural networks with attention mechanisms are used for both the word and the sentence levels. Experiments on real life datasets demonstrate effectiveness of our proposed model.

• The KIPS Transactions:PartB
• /
• v.16B no.5
• /
• pp.359-366
• /
• 2009

Analysis of 3-dimensional (3D) protein structure plays an important role of structural bioinformatics. The protein structure alignment is the main subjects of the structural bioinformatics and the most fundamental problem. Protein Structures are flexible and undergo structural changes as part of their function, and most existing protein structure comparison methods treat them as rigid bodies, which may lead to incorrect alignment. We present a new method that carries out the flexible structure alignment by means of finding SSPs(Similar Substructure Pairs) and flexible points of the protein. In order to find SSPs, we encode the coordinates of atoms in the backbone of protein into RDA(Relative Direction Angle) using local similarity of protein structure. We connect the SSPs with Floyd-Warshall algorithm and make compatible SSPs. We compare the two compatible SSPs and find optimal flexible point in the protein. On our well defined performance experiment, 68 benchmark data set is used and our method is better than three widely used methods (DALI, CE, FATCAT) in terms of alignment accuracy.

• International Journal of Highway Engineering
• /
• v.17 no.2
• /
• pp.31-37
• /
• 2015

PURPOSES : To design a pre-heater for warm in-place recycling equipment, three different heating systems were evaluated to determine their thermal efficiency. METHODS: In this study, a $30cm{\times}30cm{\times}15cm$ wheel-tracking specimen was used to measure the inner temperature as a function of the heating system. The inner temperature of the specimen was measured with a data logger at the surface, and at depths of 1cm, 2cm, 3cm, 4cm, and 5cm. To evaluate the thermal efficiency, the researchers used three different types of equipment, namely, IR, a heating wire, and a gas burner. RESULTS: The IR heating system exhibits a higher level of performance than the others to achieve the target temperature at a depth of 5cm in the specimen. The gas burner system was capable of heating the surface to a temperature of up to $600^{\circ}C$. The other types, however, cannot heat the surface up to 600. The thermal efficiencies were measured based on the laboratory conditions. CONCLUSIONS: To find the most effective system for application to the development of a pre-heater for warm in-place recycling, various systems (IR, heating wire, gas burner) were examined in the laboratory. As a result, it was found that the hot plate of a gas burner system provides the highest temperature at the surface of the asphalt but, of all the systems, the IR system provides the best internal temperature increase rate. Furthermore, a gas burner can age the asphalt binder of the surface layer as a result of the high temperature. However, the gas burner cannot attain the target temperature at 5cm. The IR system, on the other hand, is effective at increasing the internal temperature of asphalt.

• International Journal of Highway Engineering
• /
• v.13 no.1
• /
• pp.239-249
• /
• 2011

The behavior of the fiber reinforced concrete (FRC) base under environmental loads was analyzed numerically as a fundamental study to develop a high structural and functional performance composite pavement system in which the base was formed using FRC and the asphalt or cement concrete surface was placed on it. A two-dimensional finite element model of the FRC base was developed and the sensitivity study was performed with the variables including slab thickness of base, thermal expansion coefficient, elastic modulus, and tensile and compressive strengths. The crack spacing and crack width were selected as representatives of the base behavior. The effects of the selected variables on the crack spacing and crack width were analyzed and the sensitive variables were determined. The results of this study could be useful to determine the optimal material properties of the FRC base for combining well with the surface materials.

• 한국태양에너지학회:학술대회논문집
• /
• 2008.11a
• /
• pp.63-68
• /
• 2008

Efforts tp overcome the current challenge of global warming and abnormal temperature are being taken around the world. According to a report, average temperature of Korea has increased by about $0.8^{\circ}C$ for a century. In particular, temperature has rapidly increased since year 2000. Climate changes have brought remarkable changes in our lives. For example, agricultural field will see changes in crops and production. Energy used to maintain and manage architectures will be changed as well. In order to actively cope with rapidly changing global climate which drives changes from the basic behavior of our lives to subtle changes, international cooperation and researches are performed around the word. For instance, as a part of these global efforts, research on typical meterological data for computer simulation program to evaluate architecture energy performance is in progress in Korea. In order to conduct research on typical meterological data in format of data per time, reference regarding monthly maximum/minimum temperature time is required. Unfortunately, however, reference regarding maximum/minimum temperature time hasn't been defined in Korea. Therefore, this study aims to provide fundamental data essential for various researches by calculating maximum/minimum temperature time of major cities across Korea. According to the study, maximum temperature occurs at 3 p.m and minimum temperature occurs at 5 a.m or 6 a.m. respectively, in overall areas.

• Korean Journal of Optics and Photonics
• /
• v.11 no.4
• /
• pp.239-245
• /
• 2000

In this paper, the fundamental properties of diffractive optical element were investigated. Also, this work deals with theoretical approaches for achromatization in DOE's optical system based on thin lens theory. It is found that achromatization could be satisfied by one hybrid lens only, which is composed of a diffractive and a refractive element. In order to have compact optical system, we used the tele-photo type lens composed of a positive and a negative power elements instead of retro-focus lens. From the Gaussian brackets and Seidel aberration theory, the initial design was numerically obtained. The aberration properties of an initial design was aplanat and flat field. In order to correct the chromatic aberrations, refractive and diffractive elements were used on front element. This hybrid lens is also useful for correction of higher order aberrations. Compared to conventional design composed of refractive lenses only, this approach dramatically improved the compactness of the optical system. Finally, residual aberration balancing results in a lens with focal length of 3.89 mm and overall length of 5.19 mm, which has enough performance over an f-number of 4.0. Also, it is expected to fulfill all the requirements of a digital still camera lens. This optical system is superior to the current refractive lens system in the number of elements, weight, and aberration properties. rties.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.19 no.1
• /
• pp.87-97
• /
• 2015

In general, such ground based methods are utilized to validate maneuvering and attitude control logics of a spacecraft by a simulation with a flight software at a design phase and a integrated function test with actual hardwares at a system level. Recently, varification researches using operating simulators are getting increase using compact and precise components under a ground condition. The present paper investigates and summarized the development trend of cold gas propulsion systems for the spacecraft simulators and their major performance characteristics to derive fundamental data which are necessary for a conceptual design of the simulator.

• Korean Journal of Agricultural Science
• /
• v.41 no.1
• /
• pp.59-63
• /
• 2014

Precision farming aims at reduced environmental impacts with increased productivity. Soils are multi-functional media in which air, water and biota occur together and form an essential part of the landscape with a fundamental role in the environment. The requirement for herbicides and fertilizers can vary within a field in response to spatial differences in soil properties. Near infrared (NIR) spectroscopy is widely used today as a nondestructive analytical technique which is capable of determining a number of physio-chemical parameters. The objectives of this study were to develop optimal models to predict chemical properties of paddy soils by visible and NIR reflectance spectra. Total of 60 soil samples were collected in spring from 20 paddy fields within central regions in Korea. Reflectance spectra, moisture contents, pH, total nitrogen (N), organic matter, available phosphate ($P_2O_5$) of soil samples were measured. The reflectance spectra were measured in wavelength ranges of 400-2,500 nm with 2 nm interval. The method of partial least square (PLS) analysis was used to determine the soil properties. The PLS analyses showed good correlation between predicted and measured chemical properties of paddy soils in the wavelength range of 1,800-2,400 nm. Especially, it showed better performance than the previous results which used the entire wavelength range of the spectrophotometer, without considering the optimal wavelength of each soil properties.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.21 no.2
• /
• pp.143-151
• /
• 2010

In this paper, a novel rectangular patch antenna is proposed to have Zeroth Order Resonance(ZOR) generated based on the Metamaterial Complosite Right- and Left-Handed(CRLH) structure. Making the in-phase electric field over the entire antenna other than a half-wavelength as the fundamental resonance mode of a standard microstrip patch or its positive multiple, the metallic patch is suggested to be capacitively coupled with only one surrounding rectangular ring, different from the previous 1D ZOR antennas commonly having several metal cells in line. The performance of the proposed antenna is simulated by a 3D field solver that inputs the sizes of the physical structure corresponding to the equivalent circuit designed to have ZOR at 2.4 GHz. Consequently, the resonance frequency, the gain and the antenna efficiency are observed 2.4 GHz, 5 dB and 98%, respectively. Besides, the important property of the proposed antenna is addressed as the combination of the low profile as an advantage of microstrip patch antennas, and the omni-directional field pattern typical of monopole antennas.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.6 no.2
• /
• pp.288-295
• /
• 2011

The paper proposes a localization method of an underwater robot using Monte Carlo Localization(MCL) approach. Localization is one of the fundamental basics for autonomous navigation of an underwater robot. The proposed method resolves the problem of accumulation of position error which is fatal to dead reckoning method. It deals with uncertainty of the robot motion and uncertainty of sensor data in probabilistic approach. Especially, it can model the nonlinear motion transition and non Gaussian probabilistic sensor characteristics. In the paper, motion model is described using Euler angles to utilize the MCL algorithm for position estimation of an underwater robot. Motion model and sensor model are implemented and the performance of the proposed method is verified through simulation.