• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.34C no.7
• /
• pp.70-81
• /
• 1997

In this paper, we propose HMCNN(hybrid multiple component neural networks) that enhance performance of MCNN by adapting new pattern partitioning algorithm which can cluster many input patterns efficiently. Added neural network performs similar learning procedure that of kohonen network. But it dynamically determine it's number of output neurons using algorithms that decide self-organized number of clusters and patterns in a cluster. The proposed network can effectively be applied to problems of large data as well as huge networks size. As a sresutl, proposed pattern partitioning network can enhance performance results and solve weakness of MCNN like generalization capability. In addition, we can get more fast speed by performing parallel learning than that of other supervised learning networks.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1996.10a
• /
• pp.76-79
• /
• 1996

본 논무에서는 한국어 음소를 대상으로 Kohonen 이 제안한 LVQ2 방법의 결저을 보완한 MLVQ2 방법으로 인식실험을 행하고 MLVQ2 알고리즘의 유효성을 검토하고자 한다. 인식실험을 위한 음성자료는 ETRI 611단어로부터 추출한 49음소를 사용하였다. 그리고 인식실험에 있어서는 먼저 파열음을 대상으로 학습회수, 표준패턴의 수, 샘플수에 따른 인식률의 변화를 조사하였으며, 이 결과 표준패턴의 수 15개, 학습회수 10회 이하, 샘플 수 3000 개일 경우가 가장 좋은 인식률을 보였다. 이 결과를 참고로 음소군별 인식실험 결과 모음 69.11%, 파열음 74.69%, 마찰음 및 파찰음 86.31%비음 및 유음 74.51%의 평균 인식률을 얻었다. 또한 , 한국어 49음소 전음소에 대한 인식실험 결과 71.2%의 인식률 얻어 MLVQ2의 유효성을 확인하였다.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.17 no.9
• /
• pp.1006-1012
• /
• 2004

In this paper, neural networks is studied to apply as a PD source classification in XLPE power cable specimen. Two learning schemes are used to classification; BP(Back propagation algorithm), SOM(self organized map - kohonen network). As a PD source, using treeing discharge sources in the specimen, three defected models are made. And these data making use of a computer-aided discharge analyser, statistical and other discharge parameters is calculated to discrimination between different models of discharge sources. And a]so these distribution characteristics are applied to classify PD sources by two scheme of the neural networks. In conclusion, recognition efficiency of BP is superior to SOM.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.9
• /
• pp.985-996
• /
• 1990

An inverse kinematic solution of a robot manipulator using multilayer perceptrons is proposed. Neural networks allow the solution of some complex nonlinear equations such as the inverse kinematics of a robot manipulator without the need for its model. However, the back-propagation (BP) learning rule for multilayer perceptrons has the major limitation of being too slow in learning to be practical. In this paper, a new algorithm named Dynamically Reconfiguring BP is proposed to improve its learning speed. It uses a modified version of Kohonen's Self-Organizing Feature Map (SOFM) to partition the input space and for each input point, select a subset of the hidden processing elements or neurons. A subset of the original network results from these selected neuron which learns the desired mapping for this small input region. It is this selective property that accelerates convergence as well as enhances resolution. This network was used to learn the parity function and further, to solve the inverse kinematic problem of a robot manipulator. The results demonstrate faster learning than the BP network.

• Proceedings of the KIEE Conference
• /
• 1997.07c
• /
• pp.914-916
• /
• 1997

This paper presents the development of an Artificial Neural Networks(ANN) for Short-Term Load Forecasting(STLF). First, used historical load data is divided into 5 patterns for the each seasonal data using Kohonen networks. Second, classified data is used as inputs of Back-propagation networks for next day hourly load forecasting. The proposed method was tested with KEPCO hourly record (1994-95) and we obtained desirable results.

• Proceedings of the KIEE Conference
• /
• 1997.07c
• /
• pp.1130-1132
• /
• 1997

This paper proposed an application of artificial neural networks to security assessment(SA) in power system. The SA is a important factor in power system operation, but conventional techniques have not achieved the desired speed and accuracy. Since the SA problem involves classification, pattern recognition, prediction, and fast solution, it is well suited for Kohonen neural network application. Self organizing feature map(SOFM) algorithm in this paper provides two dimensional multi maps. The evaluation of this map reveals the significant security features in power system. Multi maps of multi prototype states are proposed for enhancing the versatility of SOFM neural network to various operating state.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.10a
• /
• pp.995-998
• /
• 2013

Aerial imaging is one of the most common and versatile ways of obtaining information from the Earth surface. In this paper, we present an approach by SOM(Self Organization Map) algorithm with 2 phase learning to be applied successfully to aerial images clustering due to its signal-to-noise independency. A comparison with other classical method, such as K-means and traditional SOM, of real-world areal image clustering demonstrates the efficacy of our approach.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.17 no.12
• /
• pp.3063-3072
• /
• 1993

The layout is an important and difficult problem in industrial applications like sheet metal manufacturing, garment making, circuit layout, plant layout, and land development. The module layout problem is known to be non-deterministic polynomial time complete(NP-complete). To efficiently find an optimal layout from a large number of candidate layout configuration a heuristic algorithm could be used. In recent years, a number of researchers have investigated the combinatorial optimization problems by using neural network principles such as traveling salesman problem, placement and routing in circuit design. This paper describes the application of Self-organizing Feature Maps(SOM) of the Kohonen network and Simulated Annealing Algorithm(SAA) to the layout problem of the two-dimensional rectangular shapes.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2003.05a
• /
• pp.47-51
• /
• 2003

인터넷을 통해 제공되는 맡은 양의 뉴스 정보 중에서 찾고자 하는 정확한 정보를 빠른 시간 안에 검색하고, 원하는 정보만 필터링 하는 것이 필요하다. 먼저, 인터넷에 접속된 뉴스서버들의 뉴스 문서를 각 그룹별로 수집한다. 수집된 뉴스 문서를 대상으로 퍼지추론을 통하여 문서를 대표하는 키워드를 추출하여 데이터베이스에 저장한다. 각 뉴스그룹의 문서에서 단어들을 분석하여 입력된 단어들의 개수를 이용하여 정규화 시켜서 대표적인 비지도학습 신경망인 코호넨 신경망을 사용하여 학습시킨다. 코호넨 신경망으로 추출된 단어들의 연관성을 활용하여 뉴스그룹을 클러스터링한다. 최종적으로 사용자가 관심 있는 키워드를 입력하면, 학습된 신경망이 유사한 뉴스그룹들을 사용자에게 제시해준다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04b
• /
• pp.649-651
• /
• 2002

본 논문에서는 오염된 물고기의 특성을 자동으로 분석하기 위한 진보적 행동 분석 시스템을 제안한다. 이 행동 분석 시스템은 수질 생명체들을 오염으로부터 보호할 수 있도록 하기 위한 경보 시스템으로서, 물고기의 행동 특성을 Kohonen Neural Network를 사용하여 자동으로 군집화하고 분석할 수 있도록 하였다. 이때, Neural Network의 입력으로 사용하기 위한 특징 벡터는 물고기의 좌표 위치만을 사용하지 않고 위치 좌표를 바탕으로 속도, 가속도, 각속도, 각 가속도를 구하여 이를 사용함으로써 보다 효율적인 특징 추출이 이루어질 수 있도록 하였다. 오염 생명체와 비오염 생명체의 특징을 각각 추출하여 실험해 본 결과, 오염물질에 노출된 물고기의 밤(야간) 데이터에서 다른 군집과는 다른 뚜렷한 이상 행동 특성이 나타나는 것을 알 수 있었다.