• Proceedings of the KIEE Conference
• /
• 2000.07d
• /
• pp.2696-2698
• /
• 2000

본 연구에서는 실수형 염색체들로 구성된 개체에 대해 감수분열을 적용하여 개체를 만들고, 이 생식체들의 랜덤한 선택과 교배에 의해 세대가 진화함에 따라 탐색을 수행하는 감수분열 유전알고리즘을 이용하여 퍼지모델의 최적 구조와 파라미터를 탐색하고 Gradient Descent 알고리즘으로 파라미터를 정밀 조정하는 방안을 제안한다. 제안된 방안을 적용하여 Box-Jenkins의 가스로 데이터에 대한 퍼지모델을 구성하고 그 적용 가능성을 보인다.

• Proceedings of the KIEE Conference
• /
• 2007.07a
• /
• pp.1846-1847
• /
• 2007

본 논문에서는 비선형 공정의 모델링 성능을 향상시키기 위하여 퍼지 엔트로피 분석을 통해 새로운 클러스터를 생성하고, 이 클러스터를 퍼지 모델의 새로운 규칙으로 추가하는 자기구성 퍼지 모델을 제안한다. 퍼지 엔트로피가 상대적으로 큰 데이터 집합으로 새로운 클러스터를 구성하면 퍼지 모델의 애매모호한 정도가 작아져서 모델링 오차가 줄어들 가능성이 크게 된다. 제안한 방법의 유용성을 입증하기 위해 이를 Box-Jenkins의 가스로 공정에 적용하여 퍼지 규칙수의 증가에 따른 모델링 성능의 변화를 보이고, 기존의 방법에 의한 모델링 결과와 비교한다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.3
• /
• pp.264-269
• /
• 2001

In this paper, we interpret a score as a time series and deal with the fuzzy logic-based modeling of it. The musical notes in a score represent a lot of information about the length of a sound and pitches, etc. In this paper, using melodies, tones and pitches in a score, we transform data on a score into a time series. Once more, we foml the new Lime series by sliding a window through the time series. For analyzing the time series data, we make use of the Box-Jenkins s time series analysis. On the basis of the identified characteristics of time series, we construct the fuzzy model.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.5
• /
• pp.512-519
• /
• 2003

In this paper, we propose a neuro-fuzzy modeling to improve the performance using the hierarchical clustering and Gaussian Mixture Model(GMM). The hierarchical clustering algorithm has a property of producing unique parameters for the given data because it does not use the object function to perform the clustering. After optimizing the obtained parameters using the GMM, we apply them as initial parameters for Adaptive Network-based Fuzzy Inference System. Here, the number of fuzzy rules becomes to the cluster numbers. From this, we can improve the performance index and reduce the number of rules simultaneously. The proposed method is verified by applying to a neuro-fuzzy modeling for Box-Jenkins s gas furnace data and Sugeno's nonlinear system, which yields better results than previous oiles.

• Journal of the Korea Computer Industry Society
• /
• v.4 no.12
• /
• pp.1033-1042
• /
• 2003

In this study, the economic population prediction by computer simulation has been studied by using statistical model. The forecast of future population based on that of the past is a very difficult problem as uncertain conditions are modeled in it. Even if a thought forecast is possible, world-wide cultures and the local culture emotion the cultures of the world and out country can not be predicted due to rapid change and the estimation of population is ‘nowadays more and more’ difficult to be made good guess. In the estimation of economic population, by using the census population from 1960 to 1990, and using ARIMA model developed by Box and Jenkins, the estimation has been done on the economic population until 2021 according to age as appeared table and appendix. This kind of forecast would have both good point and weak point of ARIMA model theory saying that prediction can be done only by the economic population.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.12
• /
• pp.667-674
• /
• 2000

In this paper, a modified real-valued genetic algorithm is developed by using the meiosis for human's chromosome. Unlike common crossover methods adapted in the conventional genetic algorithms, our suggested modified real-valued genetic algorithm makes gametes by conducting the meiosis for individuals composed of chromosomes, and then generates a new individual through crossovers among those. Ultimately, when appling it for the gas data of Box-Jenkin, model and parameter identifications can be concurrently done to construct the optimal model of a neural network in terms of minimizing with the structure and the error.

• Water for future
• /
• v.17 no.4
• /
• pp.281-291
• /
• 1984

In our Koreans river basins there are many of monthly rainfall data, but unfortrnately streamflow data needed are rare. Analysing monthly rainfall data of Somjin river basin, the stochastic theory model for calculation of monthly streamflow series of that region is determined. The model is composed of Box & Jenkins stansfer function plus ARIMA residual models. This linear stochastic differenced time series equation models can adapt themselves to the structure and variety of rainfall, streamflow data on the assumption of the stationary covarience. The fiexibility of Box-Jenkins method consists mainly in the iterative technique of building an AIRMA model from observations and by the use of autocorrelation functions. The best models for Somjin river basin belong to the general calss: $Y_t=($\omega$o-$\omega$_1B) C_iX_t+$\varepsilon$t$ $Y_t$ monthly streamflow, $X_t$ : monthly rainfall, $C_i$ :monthly run-off, $$\omega$o-$\omega$_1$ : transfer parameter, $$\varepsilon$_t$ : residual The streamflow series resulted from the proposed model is satisfactory comparing with the exsting streamflow data of Somjin gauging station site.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.05a
• /
• pp.245-251
• /
• 2002

본 논문에서는 상대적으로 큰 퍼지 엔트로피를 갖는 입력-출력 데이터 집단에 다중 회귀 분석을 적용하여 다차원 평면 클러스터를 생성하고, 이 클러스터를 새로운 퍼지 모델의 규칙으로 추가한 후 퍼지 모델 파라미터의 개략 동조와 정밀 동조를 수행하는 자기구성 퍼지 모델링을 제안한다. Weighted recursive least squared 알고리즘과 fuzzy C-regression model 클러스터링에 의해 퍼지 모델의 파라미터를 개략적으로 동조한 후 gradient descent 알고리즘에 의해 파라미터를 정밀 동조하면서 감수분열 유전 알고리즘을 이용하여 최적의 학습률을 탐색한다. 그리고 자기 구성 퍼지 모델링 기법을 이용하여 Box-Jenkins의 가스로 데이터, 다변수비선형 정적 함수의 데이터와 하수 처리 활성오니 공정의 모델링을 수행하고, 기존의 방법에 의한 모델링 결과와 비교하여 그 성능을 입증한다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2006.05a
• /
• pp.121-124
• /
• 2006

본 논문에서는 Bacterial Foraging Algorithm과 FCM(fuzzy c-means)클러스터링을 이용하여 TSK(Takagi-Sugeno-Kang)형태의 퍼지 규칙 생성과 퍼지 시스템(FCM-ANFIS)을 효과적으로 구축하는 방법을 제안한다. 구조동정에서는 먼저 PCA(Principal Component Analysis)을 이용하여 입력 데이터 성분간의 상관관계를 제거한 후에 FCM을 이용하여 클러스터를 생성하고 성능지표에 근거해서 타당한 클러스터의 수, 즉 퍼지 규칙의 수를 얻는다. 파라미터 동정에서는 Bacterial Foraging Algorithm을 이용하여 전제부 파라미터를 최적화 시킨다. 결론부 파라미터는 RLSE(Recursive Least Square Estimate)에 의해 추정되어진다. PCA(Principal Component Analysis)와 FCM을 적용함으로써 타당한 규칙 수를 생성하였고 Bacterial Foraging Algorithm을 이용하여 최적의 전제부 파라미터를 구하였다. 제안된 방법의 성능을 평가하기 위하여 Box-Jenkins의 가스로 데이터와 Rice taste 데이터의 모델링에 적용하였고 우수한 성능을 보임을 알 수 있었다.

• Advances in Energy Research
• /
• v.6 no.2
• /
• pp.131-143
• /
• 2019

Global Solar Radiation (GSR) is the key element for performance estimation of any Solar Power Plant (SPP). Its forecasting may help in estimation of power production from a SPP well in advance, and may also render help in optimal use of this power. Seasonal Auto-Regressive Moving Average (SARMA) and Artificial Neural Network (ANN) models are combined in order to develop a hybrid model (SARMA-ANN) conceiving the characteristics of both linear and non-linear prediction models. This developed model has been used for prediction of GSR at Gorakhpur, situated in the northern region of India. The proposed model is beneficial for the univariate forecasting. Along with this model, we have also used Auto-Regressive Moving Average (ARMA), SARMA, ANN based models for 1 - 6 day-ahead forecasting of GSR on hourly basis. It has been found that the proposed model presents least RMSE (Root Mean Square Error) and produces best forecasting results among all the models considered in the present study. As an application, the comparison between the forecasted one and the energy produced by the grid connected PV plant installed on the parking stands of the University shows the superiority of the proposed model.