• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.4
• /
• pp.355-359
• /
• 2014

In this paper, we proposed the new fuzzy pattern classifier which combines several fuzzy models with simple consequent parts hierarchically. The basic component of the proposed fuzzy pattern classifier with hierarchical structure is a fuzzy model with simple consequent part so that the complexity of the proposed fuzzy pattern classifier is not high. In order to analyze and divide the input space, we use Fuzzy C-Means clustering algorithm. In addition, we exploit Conditional Fuzzy C-Means clustering algorithm to analyze the sub space which is divided by Fuzzy C-Means clustering algorithm. At each clustered region, we apply a fuzzy model with simple consequent part and build the fuzzy pattern classifier with hierarchical structure. Because of the hierarchical structure of the proposed pattern classifier, the data distribution of the input space can be analyzed in the macroscopic point of view and the microscopic point of view. Finally, in order to evaluate the classification ability of the proposed pattern classifier, the machine learning data sets are used.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.3
• /
• pp.289-293
• /
• 2005

Pattern classification extracts various types of pattern information expressing objects in the real world and decides their class. The top priority of pattern classification technologies is to improve the performance of classification and, for this, many researches have tried various approaches for the last 40 years. Classification methods used in pattern classification include base classifier based on the probabilistic inference of patterns, decision tree, method based on distance function, neural network and clustering but they are not efficient in analyzing a large amount of multi-dimensional data. Thus, there are active researches on multiple classifier systems, which improve the performance of classification by combining problems using a number of mutually compensatory classifiers. The present study identifies problems in previous researches on multiple SVM classifiers, and proposes BORSE, a model that, based on 1:M policy in order to expand SVM to a multiple class classifier, regards each SVM output as a signal with non-linear pattern, trains the neural network for the pattern and combine the final results of classification performance.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.38 no.11
• /
• pp.810-820
• /
• 2001

Conventional binary adder requires a carry propagation to the most significant bit, and leads to serial addition. However, optical adder using a modified signed digit(MSD) number system has been Proposed to reduce the carry propagation chain encountered in binary adder. In this paper, in order to minimize the number of symbolic substitution(SS) rules, nine input patterns were divided into five groups of the same addition results. For recognizing the input reference patterns, serial connections of joint spatial encoded patterns and masks without any other spatial operations are used.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.25 no.1
• /
• pp.32-38
• /
• 2014

In this paper, we design a secant pattern antenna combined with a septum circular polarizer to be used in the satellite data communications. To generate a radiation pattern with secant type, Fourier relationship between current distribution and space factor is applied and the circular symmetric structure of the aperture simplifies the design. The optimization process is performed by combining the radiator with the septum circular polarizer, and it is confirmed that the designed radiation pattern, axial ratio, and return loss characteristics satisfies the required performance of the antenna.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.397-400
• /
• 2002

최근 패턴인식분야에서 성능향상을 목적으로 개별인식기들을 결합하거나 동적 선택에 대한 연구가 활발하게 진행되고 있다. 인식기를 동적으로 선택하는 경우에는 전체 학습패턴으로부터 학습한 개별 인식기를 이용하거나 클러스터링 알고리즘을 이용하여 학습패턴들을 특징공간에서의 부 영역으로 분할한 다음 각 클래스를 하나의 영역과 대응하는 방법이 사용되어 왔다. 이러한 접근방법에서는 각 패턴의 지역적인 정보를 이용하기 때문에 클래스 사이의 결정 경계부분에 대한 지역적인 정보를 이용하기 어렵다. 본 논문에서는 학습패턴의 지역적 영역에 대한 마이크로 인식기를 설계하여 임의의 테스트 패턴에 대한 지역적 영역에서 가장 성능이 좋은 인식기를 동적으로 선택하여 인식 성능을 향상시키는 방법을 제안한다.

• The Journal of the Korea Contents Association
• /
• v.2 no.1
• /
• pp.68-76
• /
• 2002

Clustering is representative method of components classification. But, previous clustering method that use cohesion and coupling can not be effective, because design pattern has consisted by relation between classes. In this paper, we classified design patterns with special quality of pattern structure. Classification by clustering had expressed higher correctness degree than classification by facet. Therefore, can do that it is effective that classify design patterns using clustering algorithms that is automatic classification method. When we are searching design patterns, classification of design patterns can compare and analyze similar patterns because similar patterns is saved to same category. Also we can manage repository efficiently because of using and storing link information of patterns.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.14 no.12
• /
• pp.1269-1275
• /
• 2003

In this paper, nonradiative dielectric(NRD) rotman lens with a gap-coupled unidirectional dielectric radiator(UDR) has been designed. Gap-coupled UDR is structurally suitable for NRD rotman lens. We have optimized NRD rotman lens for minimizing side-lobe, and calculated design parameters of UDR such as length of resonator and distance of gap using an equivalent circuit model of an evanescent NRD guide. Experimental prototype of UDR is fabricated and measured at the center frequency of 38 GHz. The simulated S-parameter and far-field radiation beam pattern of UDR show good agreements with measured data. Finally, total beam pattern of NRD rotman lens of multi-beam feed has been obtained using a measured pattern of UDR and array factor of NRD rotman lens. The obtained beam pattern shows remarkably suppressed side-lobe.

• The KIPS Transactions:PartD
• /
• v.16D no.1
• /
• pp.93-104
• /
• 2009

Modular implementation of a feature is a first step toward feature-oriented product line engineering, which develops and then utilizes core assets to configure products in terms of features. Aspect-oriented programming provides effective mechanisms for improving the modularity of feature implementations. However, as features in general are not independent of each other, changes in the implementation of one feature may cause changes to or side effects in the implementation of other features. Moreover, since the time at which a feature is incorporated into products, called feature binding time, may be various from compile time through load time to run time, a feature may have to be implemented differently depending on when the feature is bound into a product. To make each feature implementation module as independent as possible, this paper proposes aspectual implementation patterns that can effectively separate feature dependencies as well as feature binding time from feature implementation modules. These patterns enable flexible composition of feature implementation modules without affecting other modules according to feature selection. The approaches are demonstrated and evaluated based on a product line of scientific calculator applications.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1985.10a
• /
• pp.57-58
• /
• 1985

• KIPE Magazine
• /
• v.24 no.3
• /
• pp.82-87
• /
• 2019