• Bulletin of the Korean Chemical Society
• /
• 제26권3호
• /
• pp.399-408
• /
• 2005

We have utilized neural network (NN) studies to predict impact sensitivities of various types of explosive molecules. Two hundreds and thirty four explosive molecules have been taken from a single database, and thirty nine molecular descriptors were computed for each explosive molecule. Optimization of NN architecture has been carried out by examining seven different sets of molecular descriptors and varying the number of hidden neurons. For the optimized NN architecture, we have utilized 17 molecular descriptors which were composed of compositional and topological descriptors in an input layer, and 2 hidden neurons in a hidden layer.

• 전자공학회논문지B
• /
• 제32B권9호
• /
• pp.1137-1153
• /
• 1995

It is not yet possible to solve the optimal number of neurons in hidden layer at neural networks. However, it has been proposed and proved by experiments that there is a limit in increasing the number of neuron in hidden layer, because too much incrememt will cause instability,local minima and large error. This paper proposes a module neural controller with pattern recognition ability to solve the above trade-off problems and to obtain fast learning convergence speed. The proposed neural controller is composed of several module having Multi-layer Perrceptron(MLP). Each module have the less neurons in hidden layer, because it learns only input patterns having a similar learning directions. Experiments with six joint robot manipulator have shown the effectiveness and the feasibility of the proposed the parallel module neural controller with pattern recognition perceptron.

• 한국정보시스템학회:학술대회논문집
• /
• 한국정보시스템학회 2002년도 추계학술대회
• /
• pp.289-312
• /
• 2002

This study suggests the Link weight analysis approach to choose input variables and an integrated model to make more accurate bankruptcy prediction model. the Link weight analysis approach is a method to choose input variables to analyze each input node's link weight which is the absolute value of link weight between an input nodes and a hidden layer. There are the weak-linked neurons elimination method, the strong-linked neurons selection method in the link weight analysis approach. The Integrated Model is a combined type adapting Bagging method that uses the average value of the four models, the optimal weak-linked-neurons elimination method, optimal strong-linked neurons selection method, decision-making tree model, and MDA. As a result, the methods suggested in this study - the optimal strong-linked neurons selection method, the optimal weak-linked neurons elimination method, and the integrated model - show much higher accuracy than MDA and decision making tree model. Especially the integrated model shows much higher accuracy than MDA and decision making tree model and shows slightly higher accuracy than the optimal weak-linked neurons elimination method and the optimal strong-linked neurons selection method.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2001년도 추계학술대회 학술발표 논문집
• /
• pp.275-278
• /
• 2001

In this paper, We evaluate the learning ability of non-monotonic DBM(Deterministic Boltzmann Machine) network through numerical simulations. The simulation results show that the proposed system has higher performance than monotonic DBM network model. Non-monotonic DBM network also show an interesting result that network itself adjusts the number of hidden layer neurons. DBM network can be realized with fewer components than other neural network models. These results enhance the utilization of non-monotonic neurons in the large scale integration of neuro-chips.

• 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%.

• 한국산학기술학회논문지
• /
• 제15권3호
• /
• pp.1724-1733
• /
• 2014

본 논문에서는 비모노톤함수(non-monotone function)인 CosExp(cosine-modulated symmetric Exponential function) 함수와 모노톤함수(monotone function)인 시그모이드 함수를 캐스케이드 코릴레이션 알고리즘(Cascade Correlation algorithm)의 학습에 병행해서 사용하여 이중나선문제(two spirals problem)의 패턴인식에 어떠한 영향이 있는지 분석하고 이어서 알고리즘의 최적화를 시도한다. 첫 번째 실험에서는 알고리즘의 후보뉴런에 CosExp 함수를 그리고 출력뉴런에는 시그모이드 함수를 사용하여 나온 인식된 패턴을 분석한다. 두 번째 실험에서는 반대로 CosExp 함수를 출력뉴런에서 사용하고 시그모이드 함수를 후보뉴런에 사용하여 실험하고 결과를 분석한다. 세 번째 실험에서는 후보뉴런을 위한 8개의 풀을 구성하여 변형된 다양한 시그모이드 활성화 함수(sigmoidal activation function)를 사용하고 출력뉴런에는 CosExp함수를 사용하여 얻게 된 입력공간의 인식된 패턴을 분석한다. 네 번째 실험에서는 시그모이드 함수의 변위를 결정하는 세 개의 파라미터 값을 유전자 알고리즘을 이용하여 얻는다. 이 파라미터 값들이 적용된 시그모이드 함수들은 후보뉴런의 활성화를 위해서 사용되고 출력뉴런에는 CosExp 함수를 사용하여 실험한 최적화 된 결과를 분석한다. 이러한 알고리즘의 성능평가를 위하여 각 학습단계 마다 입력패턴공간에서 인식된 이중나선의 형태를 그래픽으로 보여준다. 최적화 과정에서 은닉뉴런(hidden neuron)의 숫자가 28에서 15로 그리고 최종적으로 12개로 줄어서 학습 알고리즘이 최적화되었음을 확인하였다.

• 한국지능시스템학회논문지
• /
• 제11권6호
• /
• pp.505-509
• /
• 2001

본 연구에서는 학습기근을 갖는 결정론적 볼츠만 머신의 은닉충 뉴런에 비단조 활성화 함수를 적요한 경위의 학습성능을 XOR 문제와 ADD 학습에 대하여 수지 시뮬레이션을 통하여분석한다. 단조 활성화함수를 사용한 경우와 비교하여 학습 수렴률, 학습안정도, 및 학습 속도에 있어서 성능이 크게 향상됨을 확인하였다. 또한 네트워크의 막전위 분포를 조사함으로서 end-cut-off 타입의 비단조 함수를 이용한 경우에 나타는 다음 층의 뉴런에 영향을 주지 않는 뉴런의 출현, 즉, 신경회로망에 있어서 은닉층 뉴런늬 수을 자율적으로 조정하는것을 확인하였따. 이것은 학습문제에 대하여 네트워크 은닉층 뉴런의 수를 명확하게 결정할수 없는 현재의 상황에 있어서는 새로운 돌파구가 될것으로 기대된다.

• Geomechanics and Engineering
• /
• 제30권3호
• /
• pp.273-280
• /
• 2022

A certain amount of random vibration excitation to subway track is caused by subway operation. This excitation is transmitted through track foundation, tunnel, soil medium, and ground building to the ground and ground structure, causing vibration. The vibration affects ground building. In this study, the results of ANSYS numerical simulation was used to establish back-propagation (BP) neural network model. Moreover, a back-propagation neural network model consisting of five input neurons, one hidden layer, 11 hidden-layer neurons, and three output neurons was used to analyze and calculate the vertical Z-vibration level of New Capital's ground buildings of Qingdao Metro phase I Project (Line M3). The Z-vibration level under different working conditions was calculated from monolithic roadbed, steel-spring floating slab roadbed, and rubber-pad floating slab roadbed under the working condition of center point of 0-100 m. The steel-spring floating slab roadbed was used in the New Capital area to monitor the subway operation vibration in this area. Comparing the monitoring and prediction results, it was found that the prediction results have a good linear relationship with lower error. The research results have good reference and guiding significance for predicting vibration caused by subway operation.

• 전자공학회논문지B
• /
• 제29B권4호
• /
• pp.58-65
• /
• 1992

In this paper new back propagation algorithm which adds neurons to hidden layer is proposed. this proposed algorithm is applied to the pattern recognition of written number coupled with back propagation algorithm through omitting redundant learning. Learning rate and recognition rate of the proposed algorithm are compared with those of the conventional back propagation algorithm and the back propagation through omitting redundant learning. The learning rate of proposed algorithm is 4 times as fast as the conventional back propagation algorithm and 2 times as fast as the back propagation through omitting redundant learning. The recognition rate is 96.2% in case of the conventional back propagation algorithm, 96.5% in case of the back propagation through omitting redundant learning and 97.4% in the proposed algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.281-286
• /
• 2005

The main objective of the research work presented in this article is to present a systematic approach for designing a multilayer feed-forward artificial neural network based self-tuning power system stabilizer (ST-ANNPSS). In order to suggest an approach for selecting the number of neurons in the hidden layer, the dynamic performance of the system with ST-ANNPSS is studied and hence compared with that of conventional PSS. Finally the effect of variation of loading condition and equivalent reactance, Xe is investigated on dynamic performance of the system with ST-ANNPSS. Investigations reveal that ANN with one hidden layer comprising nine neurons is adequate and sufficient for ST-ANNPSS. Studies show that the dynamic performance of STANNPSS is quite superior to that of conventional PSS for the loading condition different from the nominal. Also it is revealed that the performance of ST-ANNPSS is quite robust to a wide variation in loading condition.