• Journal of Korean Institute of Industrial Engineers
• /
• v.41 no.5
• /
• pp.439-446
• /
• 2015

Recently, the importance of one-class classification problem is more increasing. However, most of existing algorithms have the limitation on providing the information that effects on the prediction of the target value. Motivated by this remark, in this paper, we suggest an efficient one-class classifier using hyper-rectangles (H-RTGLs) that can be produced from intervals including observations. Specifically, we generate intervals for each feature and integrate them. For generating intervals, we consider two approaches : (i) interval merging and (ii) clustering. We evaluate the performance of the suggested methods by computing classification accuracy using area under the roc curve and compare them with other one-class classification algorithms using four datasets from UCI repository. Since H-RTGLs constructed for a given data set enable classification factors to be visible, we can discern which features effect on the classification result and extract patterns that a data set originally has.

• Journal of Information Technology and Architecture
• /
• v.9 no.1
• /
• pp.11-19
• /
• 2012

Generally, a suffix tree is an efficient data structure since it reveals the detailed internal structures of given sequences within linear time. However, it is difficult to implement a suffix tree for a large number of sequences because of memory size constraints. Therefore, in order to compare multi-mega base genomic sequence sets using suffix trees, there is a need to re-construct the suffix tree algorithms. We introduce a new method for constructing a suffix tree on secondary storage of a large number of sequences. Our algorithm divides three files, in a designated sequence, into parts, storing references to the locations of edges in hash tables. To execute experiments, we used 1,300,000 sequences around 300Mbyte in EST to generate a suffix tree on disk.

• 한국가시화정보학회:학술대회논문집
• /
• 2002.11a
• /
• pp.15-18
• /
• 2002

We propose an image based three-dimensional shape determination system. The shape, and thus the three-dimensional coordinate information of the 3-D object, is determined solely from captured images of the 3-D object from a prescribed set of viewpoints. The approach is based on the shape from silhouette (SFS) technique and the efficacy of the SFS method is tested using a sample data set. This system may be used to visualize the 3-D object efficiently, or to quickly generate initial CAD data for reverse engineering purposes. The proposed system potentially may be used in three dimensional design applications such as 3-D animation and 3-D games.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.11a
• /
• pp.457-464
• /
• 2005

The coupled time-integration method with a staggered algorithm based on computational structural dynamics (CSD), finite element method (FEM) and computational fluid dynamics (CFD) has been developed in order to demonstrate physical vibration phenomena due to dynamic aeroelastic excitations. Virtual flutter tests for the spanwise curved wing model have been effectively conducted using the present advanced computational methods with high speed parallel processing technique. In addition, the present system can simultaneously give a recorded data fie to generate virtual animation for the flutter safety test. The results for virtual flutter test are compared with the experimental data of wind tunnel test. It is shown from the results that the effect of spanwise curvature have a tendency to decrease the flutter dynamic pressure for the same flight condition.

• Language and Information
• /
• v.9 no.1
• /
• pp.19-34
• /
• 2005

Ontology is an indispensable component in intelligent and semantic processing of knowledge and information, such as in semantic web. However, ontology construction requires vast amount of data collection and arduous efforts in processing these un-structured data. This study proposed a methodology to automatically construct and generate ontologies from Sejong Electronic Dictionary. As Sejong Electronic Dictionary is structured in XML format, it can be processed automatically by computer programmed tools into an OWL(Web Ontology Language)-based ontologies as specified in W3C . This paper presents the process and concrete application of this methodology.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.115-117
• /
• 2004

Recently, mining far association rules in distributed database environments is a central problem in knowledge discovery area. While the data are located in different share-nothing machines, and each data site grows by time. Mining global frequent itemsets is hard and not efficient in large number of distributed sewen. In many distributed databases. time component(which is usually attached to transactions in database), contains meaningful time-related rules. In this paper, we design a new DTA(distributed temporal association) algorithm that combines temporal concepts inside distributed association rules. The algorithm confirms the time interval for applying association rules in distributed databases. The experiment results show that DTA can generate interesting correlation frequent itemsets related with time periods.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.83.3-83
• /
• 2001

A method to adaptively control leg-ends trajectories of a five-legged walking robot, Cepheus-2, has been developed in terms of a kind of a table look-up method. Cepheus-2 is a five-legged robot with a pentagonal body with two joints of each leg. The robot control system has a hierarchical autonomic-integrated architecture with a main computer (PC), a manager and servo modules. Being given the goals of walking by the main computer, the manager module assigns a type of leg-end trajectories of which data are described with the work space coordinates for the legs. Every servo module generates the joint angle data. In steady walking of the robot on flat floor without obstacle, two joints have to generate the assigned trajectory and five legs ...

• Journal of the Korea Society of Computer and Information
• /
• v.7 no.1
• /
• pp.54-61
• /
• 2002

In this paper, we implement the XML modeling and data modeling by the class diagram that is made by the use case after drawing out the sequence diagram. For the XML modeling, the guiding line to map the UML classes on the XML document will be presented. Then, an example to generate the ie DTD from the UML classes will be shown under the presented guiding line. In the data modeling, the transformation method from the In classes into the object-oriented database also will be proposed. Finally, we will give an example developed by the Proposed transformation method.

• Journal of Korea Multimedia Society
• /
• v.22 no.2
• /
• pp.250-259
• /
• 2019

Big data processing technology and artificial intelligence (AI) are increasingly attracting attention. Natural language processing is an important research area of artificial intelligence. In this paper, we use Korean news articles to extract topic distributions in documents and word distribution vectors in topics through LDA-based Topic Modeling. Then, we use Word2vec to vector words, and generate a weight matrix to derive the relevance SCORE considering the semantic relationship between the words. We propose a way to recommend documents in order of high score.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.12
• /
• pp.1-9
• /
• 2018

The multiple-choice multidimensional knapsack problem (MMKP) is a variant of the well known 0-1 knapsack problem, which is known as an NP-hard problem. This paper proposes a method for solving the MMKP using the integer programming-based local search (IPbLS). IPbLS is a kind of a local search and uses integer programming to generate a neighbor solution. The most important thing in IPbLS is the way to select items participating in the next integer programming step. In this paper, three ways to select items are introduced and compared on 37 well-known benchmark data instances. Experimental results shows that the method using linear programming is the best for the MMKP. It also shows that the proposed method can find the equal or better solutions than the best known solutions in 23 data instances, and the new better solutions in 13 instances.