• 지능정보연구
• /
• 제24권1호
• /
• pp.167-181
• /
• 2018

지난 10여 년간 딥러닝(Deep Learning)은 다양한 기계학습 알고리즘 중에서 많은 주목을 받아 왔다. 특히 이미지를 인식하고 분류하는데 효과적인 알고리즘으로 알려져 있는 합성곱 신경망(Convolutional Neural Network, CNN)은 여러 분야의 분류 및 예측 문제에 널리 응용되고 있다. 본 연구에서는 기계학습 연구에서 가장 어려운 예측 문제 중 하나인 주식시장 예측에 합성곱 신경망을 적용하고자 한다. 구체적으로 본 연구에서는 그래프를 입력값으로 사용하여 주식시장의 방향(상승 또는 하락)을 예측하는 이진분류기로써 합성곱 신경망을 적용하였다. 이는 그래프를 보고 주가지수가 오를 것인지 내릴 것인지에 대해 경향을 예측하는 이른바 기술적 분석가를 모방하는 기계학습 알고리즘을 개발하는 과제라 할 수 있다. 본 연구는 크게 다음의 네 단계로 수행된다. 첫 번째 단계에서는 데이터 세트를 5일 단위로 나눈다. 두 번째 단계에서는 5일 단위로 나눈 데이터에 대하여 그래프를 만든다. 세 번째 단계에서는 이전 단계에서 생성된 그래프를 사용하여 학습용과 검증용 데이터 세트를 나누고 합성곱 신경망 분류기를 학습시킨다. 네 번째 단계에서는 검증용 데이터 세트를 사용하여 다른 분류 모형들과 성과를 비교한다. 제안한 모델의 유효성을 검증하기 위해 2009년 1월부터 2017년 2월까지의 약 8년간의 KOSPI200 데이터 2,026건의 실험 데이터를 사용하였다. 실험 데이터 세트는 CCI, 모멘텀, ROC 등 한국 주식시장에서 사용하는 대표적인 기술지표 12개로 구성되었다. 결과적으로 실험 데이터 세트에 합성곱 신경망 알고리즘을 적용하였을 때 로지스틱회귀모형, 단일계층신경망, SVM과 비교하여 제안모형인 CNN이 통계적으로 유의한 수준의 예측 정확도를 나타냈다.

• 대한본초학회지
• /
• 제21권1호
• /
• pp.1-7
• /
• 2006

Objectives: To find out the effects GGT1, an antiobestic drug widely used clinics, has on the amount of feed intake, the amount of change in the body weight and the food efficiency ratio using the data from the hGHTg obese male rats. Also, to evaluate in terms of antiobestic effects, the difference between GGT1 and reductil (sibutramine), which has been approved by the FDA of the United States. Methods: We measured the change in body weight and the amount of feed intake for 8 weeks by categorizing the hGHTg obese male rats into three groups: the control group, the GGT1 group, and the reductil (RD) group. We also evaluated the antiobestic effect by calculating the food efficiency ratio, which is the increase of bodyweight divided by the amount of feed intake. Results: In case of body weight, moderate slope of the curve in the graph of GGT1 group could mean that the weight is decreasing as time flows. In case of food efficiency ratio, the p-value was 0.745 in a test for determining if an interaction exists between the group and the point of measurement, meaning that it does not exist; also, the p-value in a test for the effect of level of repetition in food efficiency ratio according to the point of measurement equaled 0.002. Conclusion: The drug-treated groups had a greater inhibitory effect in feed intake than the control group. The results showed the food efficiency ratio had a tendency to decrease. The GGT1 group in particular was under a greater effect than the RD group.

• 수산경영론집
• /
• 제15권1호
• /
• pp.58-80
• /
• 1984

The classification of islands is prerequisite for establishing a development policy to vitalize many-sided function of islands. We try to classify the 440 inhabited islands which exist in Jeon-Nam area and Kyong-Nam area by means of PCA. PCA begins with making correlation matrix of orignal variables. From this matrix we can comprehend the rough relationships between two variables. Next, we look for the eigenvalues which are roots of characteristic equation of correlation matrix. The number of eigenvalues is equal to that of original variables. We choose the largest eigenvalue λ$_1$among them and then look for the eigenvector of correlation matrix corresponding to the largest eigenvalue. Linear combination of eigenvector obtained above and original variables is namely first Principal Component (PC). Using an eigenvalue criterion(λ$\geq$ 1), we choose 3 PCs in Jeon-Nam area and 2 PCs in Kyong-Nam area. But we decide to consider only two PCs in both areas to faciliate a comparative analysis. Now, loss of information is 31.7% in Jeon-Nam area and 26.64% in Kyong-Nam area. PCs extracted by preceding procedure have characteristics as follows. The first PC relates to aggregate size of islands in case of both areas. The second PC relates to income per household, factors of agricultural production and factors of fisheries production in Jeon-Nam area, but in Kyong-Nam area it means distance from island and income per household. A classification of islands can be attained by plotting component scores of each island in graph used two PCs as axes and grouping similiar islands. 6 groups are formed in Jeon-Nam area and 5 groups in Kyong-Nam area. The result of this study in kyong-Nam area accords with prior result of study.

• Journal of Korean Neurosurgical Society
• /
• 제64권6호
• /
• pp.864-872
• /
• 2021

Objective : The aim of our study is to investigate the cytotoxic, antioxidant, and antimicrobial effects of newly synthesized boron compounds in U87MG glioblastoma cell treatment. Methods : We synthesized boron glycine monoester (BGM) and boron glycine diester (BGD) structures containing boron atoms and determined their cytotoxic activities on glioblastoma by the MTT method. The inhibitory concentration 50 (IC₅₀) value was calculated with GraphPad Prism 5.0 program. The IC₅₀ values were administered 48 hours on U87MG glioblastoma cell. Catalase (CAT), acid phosphatase (ACP) and alkaline phosphatase (ALP) enzyme activity, malondialdehyde (MDA), total glutathione (GSH), and total protein levels were detected using spectrophotometric methods. We determined the antimicrobial activities of BGM and BGD with the disc diffusion method. Results : After 48 hours of BGM and BGD application to U87MG glioblastoma cells, we found the IC₅₀ value as 6.6 mM and 26 mM, respectively. CAT and ACP enzyme activities were decreased in BGM and BGD groups. MDA which is a metabolite of lipid peroxidation was increased in both boron compounds groups. GSH level was reduced especially in BGD group. BGM and BGD have been found to be antimicrobial effects. Conclusion : Boron compounds, especially the BGM, can provide a new therapeutic approach for the treatment of glioblastoma with their anticancer, antioxidant, and antimicrobial effects.

• Journal of the Korean Wood Science and Technology
• /
• 제49권3호
• /
• pp.254-266
• /
• 2021

Using forest resources involves solving complex and diverse tasks. At the same time, one of the key goals in the field is improving the quality of forest infrastructure. This direction requires adequate mathematical and economic justification. Moreover, creating an effective infrastructure will not only increase the accessibility and usage volumes of wood and other forest resources, but also contribute to the development of continuous and sustainable forest management. The existing practice of making decisions in terms of the organizational and technological aspects of logging, based on the personal experiences of managers or leading specialists in enterprises, hinders the achievement of constant optimal efficiency. The paper presents results that are a continuation of the research cycle of the authors' team in the fields of optimization and algorithmization of various logging processes. The focus of the study lies in the processing and movement of wood resources, the most valuable products of the investigated groups of enterprises. To this end, the paper presents a developed algorithm for determining an effective technological chain of transportation in logging operations, and for improving loading and unloading processing operations under dynamic natural and production conditions. This algorithm serves as the methodological basis for designing logging infrastructure in a dynamically changing environment.

• 산업기술연구
• /
• 제33권A호
• /
• pp.31-39
• /
• 2013

There are two methods to make a distinction of deterioration of high-speed railway track. One is that an administrator checks for each attribute value of track induction data represented in graph and determines whether maintenance is needed or not. The other is that an administrator checks for monthly trend of attribute value of the corresponding section and determines whether maintenance is needed or not. But these methods have a weak point that it takes longer times to make decisions as the amount of track induction data increases. As a field of artificial intelligence, the method that a computer makes a distinction of deterioration of high-speed railway track automatically is based on machine learning. Types of machine learning algorism are classified into four type: supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. This research uses supervised learning that analogizes a separating function form training data. The method suggested in this research uses SVM classifier which is a main type of supervised learning and shows higher efficiency binary classification problem. and it grasps the difference between two groups of data and makes a distinction of deterioration of high-speed railway track.

• Journal of Ecology and Environment
• /
• 제32권2호
• /
• pp.129-135
• /
• 2009

We investigated the moth diversity on an island of southern sea of Korea. We collected moths at three sites on the island of Bogildo, Wando-gun, Jeonnam using a 22-watt ultraviolet light trap from May to October, 2008, and identified a total of 272 species and 948 individuals in 13 families. Species of Noctuidae was the most abundant, with 107 species and 318 individuals, followed by Geometridae (62 species and 147 individuals) and Pyralidae (53 species and 269 individuals). The graph of the estimated species richness in Chao 1 (432.25$\pm$37.39) did not reach an asymptote, which suggests that more moth species could be identified on the island through further sampling. An arctiid moth, Miltochrista striata, was the most abundant species captured in this study. Monthly changes in moth species richness and abundance formed M-shaped curves, with peaks in early summer (June) and late summer (August). Cluster analysis of seven sites on three islands (Aphaedo Island, Sinan-gun, Oenarodo Island, Goheung-gun and Bogildo Island) divided the sites into two groups. Distances among sites and habitat types may play an important role in determining the similarities of moth faunas among sites.

• Genomics & Informatics
• /
• 제7권3호
• /
• pp.171-174
• /
• 2009

The characteristics of metabolic pathways make them particularly amenable to layered graph drawing methods. This paper presents a visual Java-based tool for drawing and annotating biological pathways in two- and a-half dimensions (2.5D) as an alternative to three-dimensional (3D) visualizations. Such visualization allows user to display different groups of clustered nodes, in different parallel planes, and to see a detailed view of a group of objects in focus and its place in the context of the whole system. This tool is an extended version of J2dPathway.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1991년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 22-24 Oct. 1991
• /
• pp.589-594
• /
• 1991

Group Technology (GT) is a technique for identifying and bringing together related or similar components in a production process in order to take advantage of their similarities by making use of, for example, the inherent economies of flow production methods. The process of identification, from large variety and total of components, of the part families requiring similar manufacturing operations and forming the associated groups of machines is referred as 'machine-component grouping'. First part of this paper is devoted to describing a hierarchical divisive algorithm based on graph theory to find the natural part families. The objective is to form components into part families such that the degree of inter-relations is high among components within the same part family and low between components of different part families. Second part of this paper focuses on establishing cell design procedures. The aim is to create cells in which the most expensive and important machines-called key machine - have a reasonably high utilization and the machines should be allocated to minimize the intercell movement of machine loads. To fulfil the above objectives, 0-1 integer programming model is developed and the solution procedures are found. Next an attempt is made to test the feasibility of the proposed method. Several different problems appearing in the literature are chosen and the results air briefly showed.

• Korean Chemical Engineering Research
• /
• 제54권1호
• /
• pp.75-80
• /
• 2016

Biologically stabilized gold nanoparticles were synthesized from the flower aqueous extract of T. divaricata. The synthesized nanoparticles were characterized by UV-Vis spectrophotometer, Zeta sizer, FTIR and TEM analysis. T. divaricata reduced gold nanoparticles having particle size and potential of 106.532 nm and -10.2 mV, respectively, with a characteristic peak of 550 nm in UV-visible spectrophotometer. FTIR graph after comparison between the crude flower extract and gold nanoparticles showed three major shifts in the functional groups. The morphology and size of the gold nanoparticles were examined by HRTEM analysis, which showed that most of the nanoparticles were nearly spherical with size of 100 nm. The gold nanoparticles synthesized demonstrated potent anticancer activity against MCF-7 cell line. The findings conclude that the antioxidant molecule present in T. divaricata may be responsible for both reduction and capping of gold nanoparticles which possess potential applications in medicine and pharmaceutical fields.