• The Korean Journal of Applied Statistics
• /
• v.28 no.2
• /
• pp.353-360
• /
• 2015

Various statistical models have been proposed over the last decade for spatially correlated Gaussian outcomes. The spatial linear mixed model (SLMM), which incorporates a spatial effect as a random component to the linear model, is the one of the most widely used approaches in various application contexts. Employing link functions, SLMM can be naturally extended to spatial generalized linear mixed model for non-Gaussian outcomes (SGLMM). We review popular SGLMMs on non-Gaussian spatial outcomes and demonstrate their applications with available public data.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.1
• /
• pp.84-89
• /
• 2019

JPEG is a most widely used standard image compression technology. This research introduces the JPEG image compression algorithm and describes each step in the compression and decompression. Image compression is the application of data compression on digital images. The DCT (discrete cosine transform) is a technique for converting a time domain to a frequency domain. First, the image is divided into 8 by 8 pixel blocks. Second, working from top to bottom left to right, the DCT is applied to each block. Third, each block is compressed through quantization. Fourth, the matrix of compressed blocks that make up the image is stored in a greatly reduced amount of space. Finally if desired, the image is reconstructed through decompression, a process using IDCT (inverse discrete cosine transform). The purpose of this research is to review all the processes of image compression / decompression using the discrete cosine transform method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.05a
• /
• pp.107-110
• /
• 2018

Image compression is the application of Data compression on digital images. The (DCT) discrete cosine transform is a technique for converting a time domain to a frequency domain. It is widely used in image compression. First, the image is divided into 8x8 pixel blocks. Apply the DCT to each block while processing from top to bottom from left to right. Each block is compressed through quantization. The space of the compressed block array constituting the image is greatly reduced. Reconstruct the image through the IDCT. The purpose of this research is to understand compression/decompression of images using the DCT method.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.396-401
• /
• 2021

초거대 언어모델를 활용한 퓨샷(few shot) 학습법은 여러 자연어 처리 문제에서 좋은 성능을 보였다. 하지만 데이터를 활용한 추가 학습으로 문제를 추론하는 것이 아니라, 이산적인 공간에서 퓨샷 구성을 통해 문제를 정의하는 방식은 성능 향상에 한계가 존재한다. 이를 해결하기 위해 초거대 언어모델의 모수 전체가 아닌 일부를 추가 학습하거나 다른 신경망을 덧붙여 연속적인 공간에서 추론하는 P-tuning과 같은 데이터 기반 추가 학습 방법들이 등장하였다. 본 논문에서는 문맥에 따른 질의 정규화 문제를 대화형 음성 검색 서비스에 맞게 직접 정의하였고, 초거대 언어모델을 P-tuning으로 추가 학습한 경우 퓨샷 학습법 대비 정확도가 상승함을 보였다.

• Journal of the Korea Society for Simulation
• /
• v.21 no.4
• /
• pp.35-46
• /
• 2012

As complexity of real-time systems is being increased ad hoc approaches to analysis of such systems would have limitations in completeness and coverability for states space search. Formal means using a model-based approach would solve such limitations. This paper proposes a model-based formal method for logical analysis, such as safety and liveness, of real-time systems at a discrete event system level. A discrete event model for real-time systems to be analyzed is specified by DEVS(Discrete Event Systems Specification) formalism, which specifies a discrete event system in hierarchical, modular manner. Analysis of such DEVS models is performed by Communicating DEVS (C-DEVS) formalism of a timed global state transition specification and an associated analysis algorithm. The C-DEVS formalism and an associated analysis algorithm guarantees that all possible states for a given system are visited in an analysis phase. A case study of a safety analysis for a rail road crossing system illustrates the effectiveness of the proposed method of the model-based approach.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.4
• /
• pp.391-396
• /
• 2003

This paper proposes a new approach that converts continuous-valued attributes to categorical-valued ones considering the distribution of target attributes(classes). In this approach, It can be possible to get optimal interval boundaries by considering the distribution of data itself without any requirements of parameters. For each attributes, the distribution of target attributes is projected to one-dimensional space. And this space is clustered according to the criteria like as the density value of each target attributes and the amount of overlapped areas among each density values of target attributes. Clusters which are made in this ways are based on the probabilities that can predict a target attribute of instances. Therefore it has an interval boundaries that minimize a loss of information of original data. An improved performance of proposed discretization method can be validated using C4.5 algorithm and UCI Machine Learning Data Repository data sets.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.3
• /
• pp.603-611
• /
• 2013

In this paper, we propose a new method which estimates and encrypts significant component of digital image such as digital cinema using discrete wavelet packet transform (DWPT). After analyzing the characteristics of images in spatial and frequency domain, the required information for ciphering an image was extracted. Based on this information an ciphering method was proposed with wavelet transform and packetization of subbands. The proposed algorithm can encrypt images in various robust from selecting transform-level and energy threshold. From analyzing the encryption effect numerically and visually, the optimized parameter for encryption is presented. Without additional analyzing process, one can encrypt efficiently digital image using the proposed parameter. Although only 0.18% among total data is encrypted, the reconstructed image dose not identified. The paketization information of subbands and the cipher key can be used for the entire secret key.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.2
• /
• pp.165-172
• /
• 2010

In clinical data minig, choosing the optimal subset of features is such important, not only to reduce the computational complexity but also to improve the usefulness of the model constructed from the given data. Moreover the threshold values (i.e., cut-off points) of selected features are used in a clinical decision criteria of experts for differential diagnosis of diseases. In this paper, we propose a fuzzy discretization approach, which is evaluated by measuring the degree of separation of redundant attribute values in overlapping region, based on spatial distribution of data with continuous attributes. The weighted average of the redundant attribute values is then used to determine the threshold value for each feature and rough set theory is utilized to select a subset of relevant features from the overall features. To verify the validity of the proposed method, we compared experimental results, which applied to classification problem using 668 patients with a chief complaint of dyspnea, based on three discretization methods (i.e., equal-width, equal-frequency, and entropy-based) and proposed discretization method. From the experimental results, we confirm that the discretization methods with fuzzy partition give better results in two evaluation measures, average classification accuracy and G-mean, than those with hard partition.

• 전기의세계
• /
• v.39 no.3
• /
• pp.26-31
• /
• 1990

응용공학 문제들은 일반적으로 시간의 존장에 관련된 문제들이다. 해결하고자 하는 문제의 공간영역이 크고, 시간영역이 또한 큰 것이 보통인대, 이럴수록 유한요소법이나 그 밖의 다른 이산화 방법을 이용할 경우 자유도를 줄임으로서 수치해석상의 문제의 크기를 줄이는 문제는 중요한 과제가 아닐 수 없다. 고주파 성분에 대한 영향을 파악해 내기 위해서는 유한요소의 수 또는 자유도의 수를 늘려야 하므로 한없이 문제의 크기가 커질 수 있기 때문이다. 따라서 분할의 기준이 되는 오차(error)의 산정은 어떻게 구할 것인지 하는 문제 또한 중요한 연구과제가 된다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.663-666
• /
• 2021

본 논문에서는 바다와 같은 스케일이 큰 장면인 물 시뮬레이션에서 표현되는 거품 효과(Foam effects)를 노이즈 없이 디테일하게 표현할 수 있는 프레임워크를 소개한다. 거품이 생성될 위치와 거품 입자의 이류는 기존의 접근법인 스크린 투영 방법을 통해 계산한다. 이 과정에서 중요한 것이 투영맵이지만 이산화된 스크린 공간에 운동량을 투영하는 과정에서 노이즈가 발생한다. 본 논문에서는 노이즈 제거 신경망(Denoising neural network)을 활용하여 이 문제를 효율적으로 풀어낸다. 투영맵을 통해 거품이 생성될 영역이 선별되면 2D공간을 3D공간으로 역변환(Inverse transformation)하여 거품 입자를 생성한다. 결과적으로 깔끔한 거품 효과뿐만 아니라, 노이즈 제거 과정으로 인해 소실되는 거품 없이 안정적으로 거품 효과를 만들어냈다.