• 대한기계학회논문집A
• /
• 제20권2호
• /
• pp.399-407
• /
• 1996

A fuzzy learning algorithm to get the optimal fuzzy rules is presented in this paper. The algorithm introduces a reference model to generate a desired output and a performance index funtion instead of the performance index table. The performance index funtion is a cost function based on the error and error-rate between the reference and plant output. The cost function is minimized by a gradient method and the control input is also updated. In this case, the control rules which generate the desired response can be obtained by changing the portion of the error-rate in the cost funtion. In SISO(Single-Input Single- Output)plant, only by the learning delay, it is possible to experss the plant model and to get the desired control rules. In the long run, this algorithm gives us the good control rules with a minimal amount of prior informaiton about the environment.

• 한국산업정보학회논문지
• /
• 제6권1호
• /
• pp.61-69
• /
• 2001

본 연구는 관능검사에서 얻은 결과로 평가규칙을 설계하고 이를 이용하여 추후의 관능평가에 응용할 수 있는 방법을 제안함에 있어서, 퍼지추론의 규칙을 효율적으로 설계하는 것에 관련된 것이다. 퍼지추론 규칙의 수는 규칙의 전건부의 구조와 파라미터를 설계함에 있어서 퍼지분할의 수에 따라 결정되는데, 분할의 수가 많다고 해서 최적은 아니므로 효율적으로 규칙의 수를 축소하는 것이 규칙을 응용할 때의 효율성을 제고하는 동시에 실무에 응용할 때 추론엔진의 속도를 높일 수 있다. 이를 위해 본 연구에서는 선행연구에서 제시된 사례를 이용하여 추론규칙의 수를 축소하여도 대등한 결과를 얻을 수 있음을 수치예를 통하여 증명하였다. 본 연구의 결과는 향후 관능검사를 이용하는 다른 분야에도 유효하게 응용될 수 있을 것이다.

• 디지털융복합연구
• /
• 제14권1호
• /
• pp.203-209
• /
• 2016

본 논문에서는 퍼지 클러스터를 이용한 비선형 추론을 위한 퍼지 추론 시스템을 소개한다. 전형적으로, 비선형 추론을 위한 퍼지 규칙의 생성은 일반적으로 입력 벡터 차원이 증가하면 규칙의 수가 지수적으로 증가하게 된다. 이러한 문제점을 해결하기 위해, 퍼지 클러스터를 표현할 수 있는 퍼지 클러스터링 알고리즘을 이용하여 입력 벡터 공간을 분산 형태로 분할하여 퍼지 모델의 규칙을 설계한다. 이러한 방법으로 복잡하고 비선형적인 공정을 퍼지 모델링 할 수 있다. 퍼지 규칙의 전반부는 퍼지 클러스터를 갖는 FCM 클러스터링 알고리즘에 의해 결정된다. 퍼지 규칙의 후반부는 4가지 형태의 다항식 함수의 형태를 가지며, 각 규칙의 후반부 파라미터들은 표준 최소자승법을 이용함으로써 추정된다. 그리고 비선형 공정의 특성 및 성능을 평가하기 위하여 비선형 공정으로 많이 이용되고 있는 데이터를 이용한다. 실험 결과는 비선형 추론이 가능하다는 것을 보여준다.

• 전기학회논문지
• /
• 제60권1호
• /
• pp.184-192
• /
• 2011

In this study, we introduce a identification methodology for FCM-based fuzzy model. The two underlying design mechanisms of such networks involve Fuzzy C-Means (FCM) clustering method and Particle Swarm Optimization(PSO). The proposed algorithm is based on FCM clustering method for efficient processing of data and the optimization of model was carried out using PSO. The premise part of fuzzy rules does not construct as any fixed membership functions such as triangular, gaussian, ellipsoidal because we build up the premise part of fuzzy rules using FCM. As a result, the proposed model can lead to the compact architecture of network. In this study, as the consequence part of fuzzy rules, we are able to use four types of polynomials such as simplified, linear, quadratic, modified quadratic. In addition, a Weighted Least Square Estimation to estimate the coefficients of polynomials, which are the consequent parts of fuzzy model, can decouple each fuzzy rule from the other fuzzy rules. Therefore, a local learning capability and an interpretability of the proposed fuzzy model are improved. Also, the parameters of the proposed fuzzy model such as a fuzzification coefficient of FCM clustering, the number of clusters of FCM clustering, and the polynomial type of the consequent part of fuzzy rules are adjusted using PSO. The proposed model is illustrated with the use of Automobile Miles per Gallon(MPG) and Boston housing called Machine Learning dataset. A comparative analysis reveals that the proposed FCM-based fuzzy model exhibits higher accuracy and superb predictive capability in comparison to some previous models available in the literature.

• 한국정보처리학회논문지
• /
• 제6권4호
• /
• pp.988-998
• /
• 1999

본 논문에서는 LVQ(Learning vector Quantization)을 이용하여 퍼지 규칙의 수를 줄이는 방안을 제안하였다. 훈련 패턴이 많이지면 퍼지 규칙 수가 증가하게 되어 많은 기억용량과 많은 분류시간이 요구된다. 따라서 이러한 문제를 해결하기 위해서는 퍼지규칙의 수를 줄일 수 있는 방법이 강구되어야 한다. 그러나, 퍼지 규칙의 수가 줄어듦으로써 발생하는 성능의 하락을 최소화하기 위하여 양질의 초기 참조 패턴으로 훈련 한 후에, 퍼지 규칙을 생성한다. 시뮬레이션을 통해서 제안된 방법이 매우 효과적임을 알 수 있었다.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1998년도 The Third Asian Fuzzy Systems Symposium
• /
• pp.391-394
• /
• 1998

In this paper, we try to analyze two kinds of conventional neuro-fuzzy learning algorithms, which are widely used in recent fuzzy applications for tuning fuzzy rules, and give a summarization of their properties. Some of these properties show that uses of the conventional neuro-fuzzy learning algorithms are sometimes difficult or inconvenient for constructing an optimal fuzzy system model in practical fuzzy applications.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2003년도 ISIS 2003
• /
• pp.78-81
• /
• 2003

This paper proposes a fuzzy classifier system (FCS) using hyper-cone membership functions (HCMFs) and rule reduction techniques. The FCS can generate excellent rules which have the best number of rules and the best location and shape of membership functions. The HCMF is expressed by a kind of radial basis function, and its fuzzy rule can be flexibly located in input and output spaces. The rule reduction technique adopts a decreasing method by merging the two appropriate rules. We applay the FCS to a tubby rule generation for the inverted pendulum control.

• 한국감성과학회:학술대회논문집
• /
• 한국감성과학회 2000년도 춘계 학술대회 및 국제 감성공학 심포지움 논문집 Proceeding of the 2000 Spring Conference of KOSES and International Sensibility Ergonomics Symposium
• /
• pp.253-257
• /
• 2000

the objective of this study is to design and evaluate a methodology for diagnosing the appendicitis in a fuzzy neural network that integrates the partition of input space by fuzzy entropy and the generation of fuzzy control rules and learning algorithm. In particular the diagnosis of appendicitis depends on the rule of thumb of the experts such that it associates with the region, the characteristics, the degree of the ache and the potential symptoms. In this scheme the basic idea is to realize the fuzzy rle base and the process of reasoning by neural network and to make the corresponding parameters of the fuzzy control rules be adapted by back propagation learning rule. To eliminate the number of the parameters of the rules, the output of the consequences of the control rules is expressed by the network's connection weights. As a result we obtain a method for reducing the system's complexities. Through computer simulations the effectiveness of the proposed strategy is verified for the diagnosis of appendicitis.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1996년도 하계학술대회 논문집 B
• /
• pp.1021-1023
• /
• 1996

This paper proposes a design of self-tuning fuzzy rules and membership functions based on genetic algorithms. Sub-optimal fuzzy rules and membership functions are found by using genetic algorithms. Genetic algorithms are used for tuning fuzzy rules and membership functions. A arbitrary speed trajectories are selected for the reference input of the proposed methods. Experimental results show the good performance in the DC motor control system with the self-tuning fuzzy controller based on genetic algorithms.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제1권1호
• /
• pp.75-80
• /
• 2001

Rule inconsistency is an important issue that is needed to be addressed while designing efficient and optimal fuzzy rule bases. Automatic generation of fuzzy rules from data sets, using machine learning techniques, can generate a significant number of redundant and inconsistent rules. In this study we have shown that it is possible to provide a systematic approach to understand the fuzzy rule inconsistency problem by using the proposed measure called the Commonality measure. Apart from introducing this measure, this paper describes an algorithm to optimize a fuzzy rule base using it. The optimization procedure performs elimination of redundant and/or inconsistent fuzzy rules from a rule base.