• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1995년도 하계학술대회 논문집 B
• /
• pp.983-985
• /
• 1995

In this paper, we propose a Self Organizing Feature Map (SOFM) Type Neural Computation Algorithm for the Travelling Salesman Problem(TSP). The actual best solution to the TSP problem is computatinally very hard. The reason is that it has many local minim points. Until now, in neural computation field, Hopield-Tank type algorithm is widely used for the TSP. SOFM and Elastic Net algorithm are other attempts for the TSP. In order to apply SOFM type neural computation algorithms to the TSP, the object function forms a euclidean norm between two vectors. We propose a Largrangian for the above request, and induce a learning equation. Experimental results represent that feasible solutions would be taken with the proposed algorithm.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제8권2호
• /
• pp.116-120
• /
• 2008

Clustering is a method for unsupervised learning. General clustering tools have been depended on statistical methods and machine learning algorithms. One of the popular clustering algorithms based on machine learning is the self organizing map(SOM). SOM is a neural networks model for clustering. SOM and extended SOM have been used in diverse classification and clustering fields such as data mining. But, SOM has had a problem determining optimal number of clusters. In this paper, we propose an improvement of SOM using gap statistic and probability distribution. The gap statistic was introduced to estimate the number of clusters in a dataset. We use gap statistic for settling the problem of SOM. Also, in our research, weights of feature nodes are updated by probability distribution. After complete updating according to prior and posterior distributions, the weights of SOM have probability distributions for optima clustering. To verify improved performance of our work, we make experiments compared with other learning algorithms using simulation data sets.

• 한국수자원학회논문집
• /
• 제51권5호
• /
• pp.383-393
• /
• 2018

지역빈도해석은 대상 지점과 수문학적 동질성을 만족하는 주변 지점을 하나의 지역으로 보고 빈도해석을 수행하는 방법이다. 따라서 동질한 지역의 구분은 지역빈도해석에 있어서 가장 중요한 가정이라고 할 수 있다. 이에 본 연구에서는 인공신경망 기법중 하나인 자기조직화지도(self-organizing map, SOM) 기법을 활용하여 강우 지역빈도해석을 위한 동질 강수 지역을 구분하였다. 지역구분 인자로는 지형 정보와 시 단위 강우 자료를 활용하였다. 최적 SOM 지도 구성을 위해 정량적 오차와 위상관계 오차를 활용하였다. 그 결과 $7{\times}6$ 배열의 42개의 노드를 갖는 모형을 선정하였고 최종적으로 강우 지역빈도해석을 위해 6개의 군집으로 구분하였다. 동질성 검토 결과 6개의 군집 모두 동질한 지역으로 나타났으며 기존의 유사하게 구분된 지역들과 이질성 척도를 비교하였을 때 좀 더 안정적인 지역 구분결과를 나타내는 것을 확인하였다.

• 한국경영과학회지
• /
• 제27권3호
• /
• pp.75-92
• /
• 2002

This paper is concerned with proposing a new market strategy for the segmented markets of the Internet stock trading. Many companies are providing various services for customers. However, the internet stock trading market is glowing rapidly absorbing a wide variety of customers showing different tastes and demographic information, so that it is necessary for us to investigate specific strategy for the segmented markets. General strategy so far in the Internet stock trading market has been to lower transaction fee according to the market trend. As the advent of rapidly enlarging market, however, more specific strategies need to be suggested for the segmented markets. In this respect, this paper applied a self-organizing map (SOM) to 83 questionnaire data collected from the Internet stock trading market in Korea, and obtained meaningful results.

• Journal of information and communication convergence engineering
• /
• 제12권2호
• /
• pp.128-134
• /
• 2014

In this paper, we present a graphics processing unit (GPU)-based matching technique for the purpose of fast feature matching between different images. The scale invariant feature transform algorithm developed by Lowe for various feature matching applications, such as stereo vision and object recognition, is computationally intensive. To address this problem, we propose a matching technique optimized for GPUs to perform computations in less time. We optimize GPUs for fast computation of keypoints to make our system quick and efficient. The proposed method uses a self-organizing map feature matching technique to perform efficient matching between the different images. The experiments are performed on various image sets to examine the performance of the system under varying conditions, such as image rotation, scaling, and blurring. The experimental results show that the proposed algorithm outperforms the existing feature matching methods, resulting in fast feature matching due to the optimization of the GPU.

• 전자통신동향분석
• /
• 제35권4호
• /
• pp.103-114
• /
• 2020

In this study, a machine learning (ML) algorithm is analyzed and summarized as a self-organizing network (SON) realization technology that can minimize expert intervention in the planning, configuration, and optimization of mobile communication networks. First, the basic concept of the ML algorithm in which areas of the SON of this algorithm are applied, is briefly summarized. In addition, the requirements and performance metrics for ML are summarized from the SON perspective, and the ML algorithm that has hitherto been applied to an SON achieves a performance in terms of the SON performance metrics.

• The Journal of the Acoustical Society of Korea
• /
• 제13권1E호
• /
• pp.20-24
• /
• 1994

이 논문에서는 새로운 음성 인식 모델인 동적 국부 자기 조직 지도 모델과 그 학습 알고리즘을 제안한다. 동적 국부 자기 조직 지도 모델은 음성의 시간적, 공간적 왜곡을 프로그래밍 기법과 국부 자기 조직 지도로 각각 정규화 시킨다. 한국어 숫자음에 대한 실험 결과로 제안하는 모델이 예측 신경회로망 모델보다 적은 수의 연결을 갖고서도 약간 높은 인식률을 보여 효과적임을 알 수 있었다.

• 한국경영과학회지
• /
• 제28권2호
• /
• pp.105-127
• /
• 2003

In this paper, four different data mining techniques, two neural networks and two statistical modeling techniques, are compared in terms of prediction accuracy in the context of bankruptcy prediction. In business setting, how to accurately detect the condition of a firm has been an important event in the literature. In neural networks, Backpropagation (BP) network and the Kohonen self-organizing feature map, are selected and compared each other while in statistical modeling techniques, discriminant analysis and logistic regression are also performed to provide performance benchmarks for the neural network experiment. The findings suggest that the BP network is a better choice among the data mining tools compared. This paper also identified some distinctive characteristics of Kohonen self-organizing feature map.

• 한국지능시스템학회논문지
• /
• 제4권1호
• /
• pp.24-34
• /
• 1994

A RMFSOC(Reference Model Following Self-Organizing Fuzzy Logic Controller) is propose in this paper. In the RMFSOC, the refernce model is introduced, where the desired control performance can be specified by an operator of the controlled process. The self-organizing level of the RMFSOC organizes the control rules of FLC which make the process output follow the reference model output. In addition, for the use of preventing improper modifications of control rules, a complementary decission rule is induced from the possible relations between the process output and reference model output. Through a simulation study, it is shown that the robustness of the control system using the proposed RMFSOC to the set-point changes and distur bances can be greatly improved being conpared with that of the control system using the Procyk and Mamdani's SOC.

• 제어로봇시스템학회논문지
• /
• 제7권12호
• /
• pp.985-992
• /
• 2001

This paper proposes a self-organizing fuzzy modeling(SOFUM)which an create a new hyperplane shaped cluster and adjust parameters of the fuzzy model in repetition. The suggested algorithm SOFUM is composed of four steps: coarse tuning. fine tuning cluster creation and optimization of learning rates. In the coarse tuning fuzzy C-regression model(FCRM) clustering and weighted recursive least squared (WRLS) algorithm are used and in the fine tuning gradient descent algorithm is used to adjust parameters of the fuzzy model precisely. In the cluster creation, a new hyperplane shaped cluster is created by applying multiple regression to input/output data with relatively large fuzzy entropy based on parameter tunings of fuzzy model. And learning rates are optimized by utilizing meiosis-genetic algorithm in the optimization of learning rates To check the effectiveness of the suggested algorithm two examples are examined and the performance of the identified fuzzy model is demonstrated via computer simulation.