• Journal of Advanced Marine Engineering and Technology
• /
• 제21권1호
• /
• pp.66-70
• /
• 1997

In this paper a new learning algorithm for curvature smoothing and improved generalization for multi-layer neural networks is proposed. To enhance the generalization ability a constraint term of hidden neuron activations is added to the conventional output error, which gives the curvature smoothing characteristics to multi-layer neural networks. When the total cost consisted of the output error and hidden error is minimized by gradient-descent methods, the additional descent term gives not only the Hebbian learning but also the synaptic weight decay. Therefore it incorporates error back-propagation, Hebbian, and weight decay, and additional computational requirements to the standard error back-propagation is negligible. From the computer simulation of the time series prediction with Santafe competition data it is shown that the proposed learning algorithm gives much better generalization performance.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1993년도 한국자동제어학술회의논문집(국내학술편); Seoul National University, Seoul; 20-22 Oct. 1993
• /
• pp.772-777
• /
• 1993

This paper presents a neuro-fuzzy modelling method that determines chemicals dosing model based on historical operation data for effective water quality control in water treatment system and calculates automatically the amount of optimum chemicals dosing against the changes of raw water qualities and flow rate. The structure identification in the modelling by means of neuro-fuzzy reasing is performed by Genetic Algorithm(GA) and Complex Method in which the numbers of hidden layer and its hidden nodes, learning rate and connection pattern between input layer and output layer are identified. The learning network is implemented utilizing Back Propagation(BP) algorithm. The effectiveness of the proposed modelling scheme and the feasibility of the acquired neuro-fuzzy network is evaluated through computer simulation for chemicals dosing control in water treatment system.

• 한국정보통신학회논문지
• /
• 제4권1호
• /
• pp.145-154
• /
• 2000

본 논문에서는 제어대상의 사전정보가 미지인 경우의 동정 및 제어를 위하여 직교함수 신경회로망을 제안한다. 제안하는 직교함수 신경회로망은 은닉층 앞에 버퍼층을 사용하고 은닉층에는 시그모이드 함수와 시그모이드 함수의 도함수로 유도한 RBF를 이용한 직교함수를 사용하였다. 제안한 방식의 유용성을 확인하기 위하여 Narendra 모델의 동정 시뮬레이션에 의해 동정능력을 검토하였다. 또한, 제어 시스템을 구성하고 시뮬레이션 및 실험을 통하여 유용성을 확인하였다.

• International journal of advanced smart convergence
• /
• 제7권4호
• /
• pp.57-65
• /
• 2018

Artificial neural network is inspired by the biological neural network. For simplicity, in computer science, it is represented as a set of layers. Many research has been made in evaluating the number of neurons in the hidden layer but still, none was accurate. Several methods are used until now which do not provide the exact formula for calculating the number of thehidden layer as well as the number of neurons in each hidden layer. In this paper model selection approach was presented. Proposed model is based on geographical analysis of decision boundary. Proposed model selection method is useful when we know the distribution of the training data set. To evaluate the performance of the proposed method we compare it to the traditional architecture on IRIS classification problem. According to the experimental result on Iris data proposed method is turned out to be a powerful one.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권3호
• /
• pp.143-148
• /
• 2022

In this paper, we employ deep neural network (DNN) to discern Android user interaction data distribution from artificial data distribution. We utilize real Android user interaction trace dataset collected from [1] to evaluate our DNN design. In particular, we use sequential model with 4 dense hidden layers and 1 dense output layer in TensorFlow and Keras. We also deploy sigmoid activation function for a dense output layer with 1 neuron and ReLU activation function for each dense hidden layer with 32 neurons. Our evaluation shows that our DNN design fulfills high test accuracy of at least 0.9955 and low test loss of at most 0.0116 in all cases of artificial data distributions.

• Journal of Biosystems Engineering
• /
• 제37권4호
• /
• pp.271-278
• /
• 2012

Worldwide trends in animal welfare have resulted in an increased interest in individual management of sows housed in groups within hog barns. Estrus detection has been shown to be one of the greatest determinants of sow productivity. Purpose: We conducted this study to develop a method that can automatically detect the estrus state of a sow by selecting optimal texture parameters from images of a sow's pudendum and by optimizing the number of neurons in the hidden layer of an artificial neural network. Methods: Texture parameters were analyzed according to changes in a sow's pudendum in estrus such as mucus secretion and expansion. Of the texture parameters, eight gray level co-occurrence matrix (GLCM) parameters were used for image analysis. The image states were classified into ten grades for each GLCM parameter, and an artificial neural network was formed using the values for each grade as inputs to discriminate the estrus state of sows. The number of hidden layer neurons in the artificial neural network is an important parameter in neural network design. Therefore, we determined the optimal number of hidden layer units using a trial and error method while increasing the number of neurons. Results: Fifteen hidden layers were determined to be optimal for use in the artificial neural network designed in this study. Thirty images of 10 sows were used for learning, and then 30 different images of 10 sows were used for verification. Conclusions: For learning, the back propagation neural network (BPN) algorithm was used to successful estimate six texture parameters (homogeneity, angular second moment, energy, maximum probability, entropy, and GLCM correlation). Based on the verification results, homogeneity was determined to be the most important texture parameter, and resulted in an estrus detection rate of 70%.

• 한국음향학회지
• /
• 제28권1호
• /
• pp.32-43
• /
• 2009

본 논문은 자동 분할과 extreme learning machine (ELM)을 이용하여 연속 심음신호에 의한 심장질환 분류의 성능을 개선한다. 자동 분할을 위한 전처리 단계에서 비정상적인 심음신호는 심잡음 (murmur)과 클릭음 (click)을 포함하고 있기 때문에 제1음 (S1)과 제2음 (S2) 시작점 검출 결과가 부정확하거나 누락되어 기존의 심장질환 분류 시스템의 정확도를 저하시키게된다. 이러한 분할 오류에 의한 성능 저하를 감소하기 위해 S1 및 S2의 위치를 찾고, S1 및 S2의 시간 차이를 이용하여 부정확한 시작점을 교정한 다음 한 주기 심음 신호를 추출한다. 특징벡터로는 단일 주기의 심음 신호로부터 추출된 멜척도 필터뱅크 로그 에너지 계수와 포락선을 사용한다. 심장질환을 분류하기 위하여 한 개의 은닉층을 가진 ELM 알고리듬을 사용한다. 9가지 심장질환 분류 실험을 수행한 결과, 제안 방법은 81.6%의 분류 정확도를 나타내며, multi-layer perceptron(MLP), support vector machine (SVM), hidden Markov model (HMM) 중에서 가장 높은 분류 정확도를 보여준다.

• 지구물리
• /
• 제2권3호
• /
• pp.209-216
• /
• 1999

전라북도 남원군과 전라남도 곡성군을 경계로 흐르고 있는 섬진강의 남서쪽 강변에서 관로 매설에 필요한 지질정보 파악을 위한 굴절파 탐사를 실시하였다. 인라인 굴절법에 앞서 수행한 지오폰 간격 1 m, 오프셋 구간 -36∼+36 m의 워커웨이 자료로부터, 평균속도 585 m/s의 마른 토사층 하부에 평균 1,326 m/s의 속도를 갖는 젖은 사력층이 거의 수평 상태로 놓여 있으며, 그 하부에는 평균 4,218 m/s의 속도를 갖는 기반암이 분포하는 것으로 분석된다. 지오폰 간격 2 m로 획득한 총 220 m의 굴절파 측선자료를 GRM (Generalized Reciprocal Method) 방법으로 해석한 결과, 평균속도가 688 m/s, 1,473 m/s, 3,776 m/s인 3개 지층이 인지되며, 리핑이 불가능할 것으로 판단되는 기반암까지의 깊이는 숨은 층(hidden layer)의 영향에 따라 최소 1.51∼2.43 m부터 최대 2.25∼3.54 m까지로 구해진다. 이 지역 자료는 굴절법의 전형적인 문제점인 숨은 층 존재로 인하여 2번째 층의 두께를 정확히 계산하는데 어려움이 있다.

• 한국환경과학회지
• /
• 제13권7호
• /
• pp.611-618
• /
• 2004

The Neural Network Models which mathematically interpret human thought processes were applied to resolve the uncertainty of model parameters and to increase the model's output for the streamflow forecast model. In order to test and verify the flood discharge forecast model eight flood events observed at Kumho station located on the midstream of Kumho river were chosen. Six events of them were used as test data and two events for verification. In order to make an analysis the Levengerg-Marquart method was used to estimate the best parameter for the Neural Network model. The structure of the model was composed of five types of models by varying the number of hidden layers and the number of nodes of hidden layers. Moreover, a logarithmic-sigmoid varying function was used in first and second hidden layers, and a linear function was used for the output. As a result of applying Neural Networks models for the five models, the N10-6model was considered suitable when there is one hidden layer, and the Nl0-9-5model when there are two hidden layers. In addition, when all the Neural Network models were reviewed, the Nl0-9-5model, which has two hidden layers, gave the most preferable results in an actual hydro-event.

• 전기학회논문지
• /
• 제64권10호
• /
• pp.1469-1478
• /
• 2015

In this paper, The methodology of Type-2 fuzzy set-based Radial Basis Function Neural Network(T2RBFNN) is proposed for Sewage Treatment Process and the simulator is developed for application to the real-world sewage treatment plant by using the proposed model. The proposed model has robust characteristic than conventional RBFNN. architecture of network consist of three layers such as input layer, hidden layer and output layer of RBFNN, and Type-2 fuzzy set is applied to receptive field in contrast with conventional radial basis function. In addition, the connection weights of the proposed model are defined as linear polynomial function, and then are learned through Back-Propagation(BP). Type reduction is carried out by using Karnik and Mendel(KM) algorithm between hidden layer and output layer. Sewage treatment data obtained from real-world sewage treatment plant is employed to evaluate performance of the proposed model, and their results are analyzed as well as compared with those of conventional RBFNN.