• 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.

• Journal of Korean Society of Steel Construction
• /
• v.12 no.6
• /
• pp.749-758
• /
• 2000

This study presents a governing equations of bending behavior of sandwich plates with thick metal, polymer composite facings. Based on zig-zag models for through thickness deformations, the transverse shear deformation of composite facings is included. All edges of plate are assumed to be simply supported. Results of the bending analysis under lateral loads are presented for the influence of various lay up sequences of antisymmetric angle-ply laminated facings. The accuracy of the approach is ascertained by comparing solutions from the sandwich plates theory with composite facings to the laminated plates theory. Since the present analysis considers the bending stiffness of the core and also the transverse shear deformations of the laminated facings, the proposed method showed higher than that calculated according to the general laminated plates theory. The information presented might be useful to design sandwich plates structure with metal, polymer matrix composite facings.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.31 no.3
• /
• pp.125-130
• /
• 2008

Multivariate analysis is a rapidly expanding approach to data analysis. One specific technique in multivariate analysis is Principal Component Analysis (PCA). PCA is a statistical technique that linearly transform a given set of variables into a new set of composite variables. These new variables are orthogonal to each other and capture most of the information in the original variables. PCA is used to reduce the number of control points to be checked by measurement system. Therefore, the structure of the data set is simplified significantly It is also shown that eigenvectors obtained by conducting principal component analysis on the basis of the covariance matrix can be used to physically interpret the pattern of relative deformation for the points. This case study reveals the twisting deformation pattern of the underbody which is the largest mode of the total variation.

• Journal of the Korea Safety Management & Science
• /
• v.17 no.1
• /
• pp.93-101
• /
• 2015

The demand from customers on better products and systems seems to be ever increasing. To meet the demand, the systems are becoming more and more complicated in terms of both scale and functionality, thereby requiring enormous effort in the development. One bright spot of this trend is that such effort has been the driving forces of the remarkable advancement in modern systems development. On the other hand, safety issues appear to be critical in many large-scale systems such as transportation and weapon systems including high-speed trains, airplanes, ships, missiles/rockets launchers, and so on. Such systems turn out to be prone to a variety of faults and thus the resultant failure can cause disastrous accidents. For the reason, they can be referred to as safety-critical systems. The systems failure can be attributed to either random or systemic factors (or sometimes both). The objective of this paper is on how to reduce potential systemic failure in safety critical systems. To do so, a proper system design is pursued to minimize the risk of systemic failure. A focus is placed on the fact that complex systems have a lot of complicated interfaces among the system elements. To effectively handle the sources of hazards at the complicated interfaces and resultant failure, a method is developed by utilizing a design structure matrix. As a case study, the developed method is applied in the design of train control systems.

• Journal of Welding and Joining
• /
• v.33 no.6
• /
• pp.6-12
• /
• 2015

Thermal fatigue life of the automobile exhaust manifold is directly affected by the restraint force according to the structure of exhaust system and bead shape of the welded joints. In the present study, the microstructural changes and precipitation behavior during thermal fatigue cycle of the 18wt% Cr ferritic stainless steel weld heat affected zone (HAZ) considering restraint stress were investigated. The simulation of weld HAZ and thermal fatigue test were carried out using a metal thermal cycle simulator under complete constraint force in the static jig. The change of the restraint stress on the weld HAZ was simulated by changing the shape of notch in the specimen considering the stress concentration factor. Thermal fatigue properties of the weld HAZ were deteriorated during cyclic heating and cooling in the temperature range of $200^{\circ}C$ to $900^{\circ}C$ due to the decrease of Nb content in solid solution and coarsening of MX type precipitates, laves phase, $M_6C$ with coarsening of grain and softening of the matrix. As the restraint stress on the specimen increased, the thermal fatigue life was decreased by dynamic precipitation and rapid coarsening of the precipitates.

• Journal of Broadcast Engineering
• /
• v.8 no.3
• /
• pp.278-287
• /
• 2003

Noise in an image degrades image quality and deteriorates coding efficiency. Recently, various edge-preserving noise filtering methods based on the nonstationary image model have been proposed to overcome this problem. In most conventional nonstationary image models, however, pixels are assumed to be uncorrelated to each other in order not to Increase the computational burden too much. As a result, some detailed information is lost in the filtered results. In this paper, we propose a computationally feasible adaptive noise smoothing algorithm which considers the nonstationary correlation characteristics of images. We assume that an image has a nonstationary mean and can be segmented into subimages which have individually different stationary correlations. Taking advantage of the special structure of the covariance matrix that results from the proposed image model, we derive a computationally efficient FFT-based adaptive linear minimum mean-square-error filter. Justification for the proposed image model is presented and effectiveness of the proposed algorithm is demonstrated experimentally.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.30 no.4
• /
• pp.316-322
• /
• 2004

A low molecular weight component named bone morphogenetic protein(BMP) chemically isolated from the organic matrix of bone, induce postfetal connective tissue cells surrounding small blood vessels to differentiate into cartilage and bone. The end product of BMP is a spherical ossicle of lamella bone filled with red bone marrow for the functional loading. This is a important point that the graft material is embedded the defect site during the implant surgery. Because present knowledge of the relationship between BMP and bone regeneration arises mainly from studies of induced bone formation in heterotopic sites, it would be helpful to determine whether BMP plays any part in the process of bone healing. The BMPs have been shown to play crucial roles in normal skeletal development as well as bone healing and are able to activate transcription of genes involved in cellular migration, proliferation, and differentiation. The delivery of BMP on matrices has been efficacious in the treatment of defect bone in implant surgery. The purpose of the histologic study was to evaluate the effect of DLB(demineralized lyophilized bone) coated with purified human BMP(hBMP-I) in immediate implant surgery with bony defect to obtain the functional structure of implant asap. The ability of a graft of hBMP-I to accelerate bony defect repair provides a rationale for its use in immediate implant surgery that have large bone defect in edentulous area.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.2
• /
• pp.531-535
• /
• 2015

Vasculogenic mimicry (VM) refers to the unique ability of highly aggressive tumor cells to mimic the pattern of embryonic vasculogenesis, which was associated with invasion and metastasis. The grape seed proanthocyanidins (GSPs) had attracted much attention as a potential bioactive anti-carcinogenic agent. However, GSPs regulation of VM and its possible mechanisms in a triple-negative breast cancer cells (TNBCs) remain not clear. Therefore, we examined the effect of GSPs on VM information in HCC1937 cell model. In this study, we identified the VM structure via the three-dimensional (3D) matrix in vitro. Cell viability was measured using the CCK8 assay. The effects of GSPs on human triple-negative breast cancer cells (TNBCs) HCC1937 in terms of related proteins of VM information were determined using western blot analysis. In vitro, the tubular networks were found in highly invasive HCC1937 cells but not in the non-invasive MCF-7 cells when plated on matrigel. The number of vascular channels was significantly reduced when cells were exposed in GSPs ($100{\mu}g$/ml) and GSPs ($200{\mu}g/mL$) groups (all p<0.001). Furthermore, we found that treatment with GSPs promoted transition of the mesenchymal state to the epithelial state in HCC1937 cells as well as reducing the expression of Twist1 protein, a master EMT regulator.GSPs has the ability to inhibit VM information by the suppression of Twist1 protein that could be related to the reversal of epithelial-to-mesenchymal (EMT) process. It is firstly concluded that GSPs may be an p otential anti-VM botanical agent for human TNBCs.

• The Journal of the Acoustical Society of Korea
• /
• v.11 no.2
• /
• pp.59-67
• /
• 1992

Recently, the problem which actively control the unwanted noise propagated from the technical structure by the generated secondary sound has become considerable topic from the environmental preservation point of view. In most of these studies, active noise control deals with a plane wave propagation at low frequency using adaptive filtering techniques. On the other hand, in real acoustic systems are mostly short due to the limitation of geometric configuration. In this case, the acoustic properties such as reflections and resonances inside the acoustic system should be considered. In this paper, the acoustic modeling method for short length duct was introduced using the transfer matrix method, and the active noise control problem was investigated with \implementation of FIR filter for the transfer function of control system derived from this modeling method. The identification methods for the acoustic model of actual control system was proposed by numerical computation technique based on the estimation of optimal FIR filter coefficients. The acceptable attenuation on the real acoustic system and stability of the controller are predicted in this computational simulation.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.5
• /
• pp.30-38
• /
• 2011

This paper proposes a new enhancing document clustering method using a synonym of cluster topic and the similarity. The proposed method can well represent the inherent structure of document cluster set by means of selecting terms of cluster topic based on the semantic features by NMF. It can solve the problem of "bags of words" by using of expanding the terms of cluster topics which uses the synonyms of WordNet. Also, it can improve the quality of document clustering which uses the cosine similarity between the expanded cluster topic terms and document set to well cluster document with respect to the appropriation cluster. The experimental results demonstrate that the proposed method achieves better performance than other document clustering methods.