• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2004.05b
• /
• pp.158-161
• /
• 2004

RBF 네트워크의 중간층은 클러스터링하는 층이다. 즉, 이 충의 목적은 주어진 자료 집합을 유사한 클러스터들(homogenous cluster)로 분류하는 것이다. 여기서 유사하다는 것은 입력 데이터들에 대한 특징 벡터 공간사이에서 한 클러스터내의 벡터들 간에 거리를 측정하여 정해진 반경 내에 존재하면 같은 클러스터로 분류하고 정해진 반경 내에 존재하지 않으면 다른 클러스터로 분류한다. 그러나 정해진 반경 내에서 클러스터링하는 것은 잘못된 클러스터를 선택하는 단점을 가지게 된다. 그러므로 중간층을 결정하는 .것은 RBF 네트워크의 전반적인 효율성에 큰 영향을 준다. 따라서 본 논문에서는 효율적으로 중간층을 결정하기 위한 방법으로 퍼지 C-Means 클러스터링 알고리즘을 적용한 퍼지 RBF 네트워크를 제안한다. 제안된 퍼지 RBF 네트워크의 학습은 크게 두 단계로 구분된다. 첫 번째 단계는 입력층과 중간층 사이에 퍼지 C-Means 알고리즘이 수행되고, 두 번째 단계는 중간층과 출력층 사이에 지도학습이 수행된다. 제안된 방법의 학습 성능을 평가하기 위하여 실제 주민등록증에서 추출한 숫자패턴에 적용한 결과, 기존의 RBF네트워크 보다 학습 성능이 개선된 것을 확인하였다.

• Proceedings of the KIEE Conference
• /
• 2005.07d
• /
• pp.2888-2890
• /
• 2005

In this paper, we introduce a new structure of fuzzy-neural networks Fuzzy Set-based Polynomial Neural Networks (FSPNN). The two underlying design mechanisms of such networks involve genetic optimization and information granulation. The resulting constructs are Fuzzy Polynomial Neural Networks (FPNN) with fuzzy set-based polynomial neurons (FSPNs) regarded as their generic processing elements. First, we introduce a comprehensive design methodology (viz. a genetic optimization using Genetic Algorithms) to determine the optimal structure of the FSPNNs. This methodology hinges on the extended Group Method of Data Handling (GMDH) and fuzzy set-based rules. It concerns FSPNN-related parameters such as the number of input variables, the order of the polynomial, the number of membership functions, and a collection of a specific subset of input variables realized through the mechanism of genetic optimization. Second, the fuzzy rules used in the networks exploit the notion of information granules defined over systems variables and formed through the process of information granulation. This granulation is realized with the aid of the hard C- Means clustering (HCM). The performance of the network is quantified through experimentation in which we use a number of modeling benchmarks already experimented with in the realm of fuzzy or neurofuzzy modeling.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.1
• /
• pp.225-231
• /
• 2013

In this study, we propose the design of optimized pRBFNNs-based night vision face recognition system using PCA algorithm. It is difficalt to obtain images using CCD camera due to low brightness under surround condition without lighting. The quality of the images distorted by low illuminance is improved by using night vision camera and histogram equalization. Ada-Boost algorithm also is used for the detection of face image between face and non-face image area. The dimension of the obtained image data is reduced to low dimension using PCA method. Also we introduce the pRBFNNs as recognition module. The proposed pRBFNNs consists of three functional modules such as the condition part, the conclusion part, and the inference part. In the condition part of fuzzy rules, input space is partitioned by using Fuzzy C-Means clustering. In the conclusion part of rules, the connection weights of pRBFNNs is represented as three kinds of polynomials such as linear, quadratic, and modified quadratic. The essential design parameters of the networks are optimized by means of Differential Evolution.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.1
• /
• pp.128-135
• /
• 2015

In this study, we develop the very short-term precipitation forecasting model as well as classifier based on polynomial radial basis function neural networks by using AWS(Automatic Weather Station) and KLAPS(Korea Local Analysis and Prediction System) meteorological data. The polynomial-based radial basis function neural networks is designed to realize precipitation forecasting model as well as classifier. The structure of the proposed RBFNNs consists of three modules such as condition, conclusion, and inference phase. The input space of the condition phase is divided by using Fuzzy C-means(FCM) and the local area of the conclusion phase is represented as four types of polynomial functions. The coefficients of connection weights are estimated by weighted least square estimation(WLSE) for modeling as well as least square estimation(LSE) method for classifier. The final output of the inference phase is obtained through fuzzy inference method. The essential parameters of the proposed model and classifier such ad input variable, polynomial order type, the number of rules, and fuzzification coefficient are optimized by means of Particle Swarm Optimization(PSO) and Differential Evolution(DE). The performance of the proposed precipitation forecasting system is evaluated by using KLAPS meteorological data.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.2
• /
• pp.416-424
• /
• 2017

In this study, we propose a design of digits recognition system based on RBFNNs through a comparative study of pre-processing algorithms in order to recognize digits in handwritten. Histogram of Oriented Gradient(HOG) is used to get the features of digits in the proposed digits recognition system. In the pre-processing part, a dimensional reduction is executed by using Principal Component Analysis(PCA) and (2D)2PCA which are widely adopted methods in order to minimize a loss of the information during the reduction process of feature space. Also, The architecture of radial basis function neural networks consists of three functional modules such as condition, conclusion, and inference part. In the condition part, the input space is partitioned with the use of fuzzy clustering realized by means of the Fuzzy C-Means algorithm. Also, it is used instead of gaussian function to consider the characteristic of input data. In the conclusion part, the connection weights are used as the extended type of polynomial expression such as constant, linear, quadratic and modified quadratic. By using MNIST handwritten digit benchmarking database, experimental results show the effectiveness and efficiency of proposed digit recognition system when compared with other studies.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.3
• /
• pp.354-359
• /
• 2008

Clustering is a method to classify the given data set with same property into several classes. To cluster data, many methods such as K-Means, Fuzzy C-Means(FCM), Mountain Method(MM), and etc, have been proposed and used. But the clustering results of conventional methods are sensitively influenced by initial values given for clustering in each method. Especially, FCM is very sensitive to noisy data, and cluster center distortion phenomenon is occurred because the method dose clustering through minimization of within-clusters variance. In this paper, we propose a clustering method which reduces cluster center distortion through merging the nearest data based on the data weight, and not being influenced by initial values. We show the effectiveness of the proposed through experimental results applied it to various types of data sets, and comparison of cluster centers with those of FCM.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.2
• /
• pp.124-132
• /
• 2014

In this paper, we present a novel backlit region detection and contrast enhancement method using fuzzy C-means clustering and adaptively partitioned block based contrast stretching. The proposed method separates an image into both dark backlit and bright background regions using adaptively partitioned blocks based on the optimal threshold value computed by fuzzy logic. The detected block-wise backlit region is refined using the guided filter for removing block artifacts. Contrast stretching algorithm is then applied to adaptively enhance the detected backlit region. Experimental results show that the proposed method can successfully detect the backlit region without a complicated segmentation algorithm and enhance the object information in the backlit region.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.4
• /
• pp.435-442
• /
• 2007

To reduce damage from high concentration ozone in the air, we have researched how to predict high concentration ozone before it occurs. High concentration ozone is a rare event and its reaction mechanism has nonlinearities and complexities. In this paper, we have tried to apply and consider as many methods as we could. We clustered the data using the fuzzy c-mean method and took a rejection sampling to fill in the missing and abnormal data. Next, correlations of the input component and output ozone concentration were calculated to transform more correlated components by modified log transformation. Then, we made the prediction models using Dynamic Polynomial Neural Networks. To select the optimal model, we adopted a minimum bias criterion. Finally, to evaluate suggested models, we compared the two models. One model was trained and tested by the transformed data and the other was not. We concluded that the modified transformation effected good to ideal performance In some evaluations. In particular, the data were related to seasonal characteristics or its variation trends.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.426-431
• /
• 1998

We propose a fuzzy rule based method for structure preserving dimensionality reduction. This method selects a small representative sample and applies Sammon's method to project it. The input data points are then augmented by the corresponding projected(output) data points. The augmented data set thus obtained is clustered with the fuzzy c-means(FCM) clustering algorithm. Each cluster is then translated into a fuzzy rule for projection. Our rule based system is computationally very efficient compared to Sammon's method and is quite effective to project new points, i.e., it has good predictability.

• Journal of the Korea Society of Computer and Information
• /
• v.20 no.12
• /
• pp.15-20
• /
• 2015

In this paper, we propose a modified fuzzy clustering algorithm which can overcome the center deviation due to the Euclidean distance commonly used in fuzzy clustering. Among fuzzy clustering methods, Fuzzy C-Means (FCM) is the most well-known clustering algorithm and has been widely applied to various problems successfully. In FCM, however, cluster centers tend leaning to high density clusters because the Euclidean distance measure forces high density cluster to make more contribution to clustering result. Proposed is an enhanced algorithm which modifies the objective function of FCM by adding a center-scattering term to make centers not to be close due to the cluster density. The proposed method converges more to real centers with small number of iterations compared to FCM. All the strengths can be verified with experimental results.