• Journal of Korean Society for Geospatial Information Science
• /
• v.8 no.2 s.16
• /
• pp.47-60
• /
• 2000

On this study, Among the classification methods of land cover using satellite imagery, we compared the classification accuracy of Neural Network Classifier and that of Maximum Likelihood Classifier which has the characteristics of parametric and non-parametric classification method. In the assessment of classification accuracy, we analyzed the classification accuracy about testing area as well as training area that many analysts use generally when assess the classification accuracy. As a result, Neural Network Classifier is superior to Maximum Likelihood Classifier as much as 3% in the classification of training data. When ground reference data is used, we could get poor result from both of classification methods, but we could reach conclusion that the classification result of Neural Network Classifier is superior to the classification result of Maximum Likelihood Classifier as much as 10%.

• Management & Information Systems Review
• /
• v.39 no.1
• /
• pp.35-57
• /
• 2020

Recently, research F&E(Facilities and Equipment) have become very important as tools and means to lead the development of science and technology. The government has been continuously expanding investment budgets for R&D and research F&E, and the need for efficient operation and systematic management of research F&E built up nationwide has increased. In December 2010, The government developed and completed a standardized classification system for national research F&E. However, accuracy and trust of information classification are suspected because information is collected by a method in which a user(researcher) directly selects and registers a classification code in NTIS. Therefore, in the study, we analyzed linearly using linear discriminant analysis(LDA) and analysis of variance(ANOVA), to measure the classification accuracy for the standardized classification system(8 major-classes, 54 sub-classes, 410 small-classes) of the national research facilities and equipment established in 2010, and revised in 2015. For the analysis, we collected and used the information data(50,271 cases) cumulatively registered in NTIS(National Science and Technology Service) for the past 10 years. This is the first case of scientifically verifying the standardized classification system of the national research facilities and equipment, which is based on information of similar classification systems and a few expert reviews in the in-outside of the country. As a result of this study, the discriminant accuracy of major-classes organized hierarchically by sub-classes and small-classes was 92.2 %, which was very high. However, in post hoc verification through analysis of variance, the discrimination power of two classes out of eight major-classes was rather low. It is expected that the standardized classification system of the national research facilities and equipment will be improved through this study.

• Journal of Korea Technology Innovation Society
• /
• v.10 no.3
• /
• pp.531-554
• /
• 2007

This paper suggests a new model connecting various knowledge activities through classification systems such as classifications of research, technology, industry, major and job. Although research activities are linked to technology and industry areas or to education and job areas, there is no effort to link these kinds of activities. There are a few studies to link research and technology or research and education respectively. But, there have been no studies to connect technology-industry linkage and education-job linkage. This paper suggests that research area can be a basis of link between technology-industry linkage and education-job linkage. The methods building the links are not simple, but easy; 1) setting up new science/research classification system having two dimensions of research and application, 2) building electronic systems and databases allowing fields for several classification systems, and 3) making rules using multi-dimensional classification systems following the purpose of the programs. The model is designed to meet the needs of nationwide R&D and human resources policies, and for the preparation of knowledge society to grasp the relationship between sequential activities using knowledge. If we know the interactive relationships between various areas, we can trace related phenomena in different activities with restricted information.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.2
• /
• pp.146-155
• /
• 2008

To improve the performance of tiled display system, two parallel tiled display algorithms are proposed by combining the sort-first and the sort-last sorting methods. In the proposed algorithms, the view frustum culling is employed along with the OpenGL display list for the sort-first sorting, and the pre-detection sort-last sparse sorting method is used for sort-last sorting. Through the benchmarking tests, the performances of two proposed algorithms are investigated. Based on the observations, it is suggested how to select an optimal algorithm among the two proposed parallel tiled display algorithms for the given visualization model.

• Korean Journal of Medicinal Crop Science
• /
• v.13 no.1
• /
• pp.21-29
• /
• 2005

This study was done to obtain their morphological traits to analyse genetic diversity and intraspecific relationship of 47 Polygonatum species collected from Gyeongnam province. Plant height was the highest in P. thunbergii but the shortest in P. involucratum. Growth habit and its colors were classified to 3 groups, respectively. Leaf shapes were sorted to 5 groups including lanceolate with petiole or none, petiole colors were done to 3 groups including a species having dark green leaves of purple colored margin. Flower shapes were divided as 3 groups of urceolate, tubular and gourd shapes, and its colors were white, greenish white and light green, especially light green in a species with gourd shape. Filament shapes were two types of flatness and cylinder. Peduncle color and bract attached below it showed 4 types, respectively. Fruit shapes were sorted to 3 groups. In 100-fruit weights P. ordoratum var. pluriflorum showed the greatest but P. involucratum did the least. Two species were completely resistant to leaf brightness although 7 species showed less than 7 % infection rates. Rhizome yields ranged from 4.4 g to 94.8 g per plant, showing their significant variation. In correlation analysis between 9 major characters, rhizome yield per plant was positively correlated with plant height, stalk diameter, leaf number, leaf length and width, and rhizome diameter but leaf brightness was negatively done with plant height, stalk diameter, leaf number and length, 100-seed weight, rhizome yield per plan and rhizome diameter.

• Korean Journal of Plant Resources
• /
• v.29 no.4
• /
• pp.467-478
• /
• 2016

This study was carried out to elucidate the distribution of vascular plants and their usefulness of Mt. Janggunbong (1,136 m) in Bonghwa-gun, Gyeoungsangbuk-do. The vascular plants, collected 15 times from 2006 to 2015, consisted a total of 462 taxa; 82 families, 279 genera, 397 species, 2 subspecies, 55 varieties, 8 forms. 10 taxa of the Korean endemic plants were recorded and 1 taxon of Critically Endangered Species (CR), 5 taxa of Vulnerable Species (VU) and 7 taxa of Least Concerned Species (LC), categorized by the Korean Forest Service as rare plants, were investigated in this region. Furthermore, Ⅳ, Ⅲ degrees of floristic regional indicator plants, designated by the Korean Ministry of Environment, were included 8 taxa and 14 taxa, respectively. Based on the usefulness, edible, pasturing, medicinal, ornamental, timber, stain, industrial, fiber and unknown usefulness plants included 352 taxa, 107 taxa, 71 taxa, 18 taxa, 8 taxa, 5 taxa, 3 taxa, 2 taxa and 111 taxa, respectively. In addition, 28 taxa of naturalized plants were observed.

• Management & Information Systems Review
• /
• v.39 no.3
• /
• pp.1-16
• /
• 2020

The current tariff return system requires tax officials to calculate tax amount by themselves and pay the tax amount on their own responsibility. In other words, in principle, the duty and responsibility of reporting payment system are imposed only on the taxee who is required to calculate and pay the tax accurately. In case the tax payment system fails to fulfill the duty and responsibility, the additional tax is imposed on the taxee by collecting the tax shortfall and imposing the tax deduction on For this reason, item classifications, together with tariff assessments, are the most difficult and could pose a significant risk to entities if they are misclassified. For this reason, import reports are consigned to customs officials, who are customs experts, while paying a substantial fee. The purpose of this study is to classify HS items to be reported upon import declaration and to indicate HS codes to be recorded on import declaration. HS items were classified using the attached image in the case of item classification based on the case of the classification of items by the Korea Customs Service for classification of HS items. For image classification, CNN was used as a deep learning algorithm commonly used for image recognition and Vgg16, Vgg19, ResNet50 and Inception-V3 models were used among CNN models. To improve classification accuracy, two datasets were created. Dataset1 selected five types with the most HS code images, and Dataset2 was tested by dividing them into five types with 87 Chapter, the most among HS code 2 units. The classification accuracy was highest when HS item classification was performed by learning with dual database2, the corresponding model was Inception-V3, and the ResNet50 had the lowest classification accuracy. The study identified the possibility of HS item classification based on the first item image registered in the item classification determination case, and the second point of this study is that HS item classification, which has not been attempted before, was attempted through the CNN model.

• Annual Conference on Human and Language Technology
• /
• 1997.10a
• /
• pp.129-135
• /
• 1997

한국어 명사들을 의미별로 분류하여 계층화시킨 '한국어 명사 의미 계층 구조'는, 한국어 문장을 처리할 때 한국어의 의미 정보를 제공할 수 있는 매우 중요한 정보들 중의 하나이다. 본 논문에서는, 국어 사전의 명사에 대한 뜻풀이말을 이용하여 bottom-up 방식으로 '한국어 명사 의미 계층 구조'를 구축하였다. 본 논문에서 구축한 '한국어 명사 의미 계층 구조'는, tree가 43개, node가 12,833개, terminal node가 10,347 개이며, 깊이가 17인 하나의 forest이다. 이것의 제 1, 2 계층(level 1,2)에서의 분류 형태는 top-down 방식에 의한 기존의 분류들과 매우 다른 모습인 반면에, 제 3 계층 이하에서의 분류 형태는 의미소성(意味素性)에 의한 기존의 분류와 거의 일치하는 모습을 나타낸다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2010.05a
• /
• pp.1380-1384
• /
• 2010

현재 환경부에서는 수질오염총량관리제를 위하여 각 단위유역의 말단지점에서 8일 간격으로 수질 및 유량을 측정하고 있으며, 이 자료들을 공개하고 있다. 이러한 양질의 자료의 활용성을 제고하기 위해서는 무엇보다도 자료의 분석을 위한 다양한 기법이 개발되고 제안되어야 한다. 따라서 본 연구에서는 수질 및 유량자료를 동시에 적용하여 두 자료 사이의 관계를 조사하고 특성을 파악하기 위하여 자기조직화 특성지도(Self-Organizing Feature Map: SOFM) 이론을 적용하였다. 시행착오법에 의해 적정한 SOFM 구조를 결정하였으며, 그 결과 $4{\times}4$ 구조의 육각형 배열을 갖는 구조를 이용하였다. SOFM에 의해 분류된 3개의 패턴 중 패턴-1은 유량자료의 크기에 의해 분류되었고, 패턴-2와 패턴-3은 BOD 농도의 크기에 따라 분류된 것으로 파악되었다. 따라서 SOFM의 적용에 의한 자료의 분류를 수행하고, 그 분류기준을 파악할 경우 SOFM의 자료 분석 도구로서의 활용성이 더욱 높아질 것으로 판단된다.

• Journal of KIISE:Software and Applications
• /
• v.30 no.5_6
• /
• pp.466-475
• /
• 2003

To build an efficient Question-Answering (QA) system, a question type classifier is needed. It can classify user's queries into predefined categories regardless of the surface form of a question. In this paper, we propose a question type classifier using a Support Vector Machine (SVM). The question type classifier first extracts features like lexical forms, part of speech and semantic markers from a user's question. The system uses $X^2$ statistic to select important features. Selected features are represented as a vector. Finally, a SVM categorizes questions into predefined categories according to the extracted features. In the experiment, the proposed system accomplished 86.4% accuracy The system precisely classifies question type without using any rules like lexico-syntactic patterns. Therefore, the system is robust and easily portable to other domains.