The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
The present paper describes an algorithmic technique to speed up weight convergence in neural networks on-line training. Standard pattern backpropagation is modified to train the neural network over a time window of samples and not one sample only, so that a faster weight convergence may be achieved. The use of such training technique is explained in an adaptive control task and problems related to...
There has been developed many method for the better convergence and generalization ability of neural network. Multilayer Perceptron (MLP) is made multi hidden layered structure for better performance. But in these types of structures still error from any output classes propagates in the backward direction which has a negative impact on the weight updating as well as overall performance because every...
In human brain the neurons are excited in a dynamic way. The response of different neurons varies widely because of the variation of electrical signal in every neuron. Backpropagation(BP) is a training algorithm where the learning of the Neural Network (NN) is done by a constant learning rate (LR). But to replicate the human brain function, the learning rate should be changed as the excitation of...
This paper proposes a hybrid algorithm for training a feed-forward neural network by combining both Particle Swarm Optimization (PSO) and Information Gain with Backpropagation (BP) algorithm. A conventional neural network training algorithm, i.e. BP, has several drawbacks in its slow convergence and local optima. Although PSO can be applied to search for the near optimal set of weights in the neural...
Backpropagation algorithm is widely used to solve many real-world problems, using the concept of Multilayer Perceptron. However, main disadvantages of Backpropagation are the convergence rate of it being relatively slow, and it is often trapped in the local minima. To solve this problem, it is found in the literatures, an evolutionary algorithm such as Particle Swarm Optimization algorithm is applied...
This paper uses generalized congruence function instead of transfer function of classical BP neural network, and improve convergence rate of neural network. We introduce the subsection generalized derivation, error back propagation derivation mechanism of classical BP algorithm to adjust weight vector in generalized congruence neural network, and modify generalized congruence neural network, and then...
After studying the disadvantage of BP neural network which has low convergent speed and trap into local minima easily, an idea of designing a new hybrid neural network model. By using Artificial Bee Colony Algorithm (ABC) to expand the updated space of weight and using the fitness functions to decide the better weight. On the basis, make the acquired better value as the weight of BP neural network...
Local minimum is incorporated problem in neural network (NN) training. To alleviate this problem, a modification of standard backpropagation (BP) algorithm, called BPCL for training NN is proposed. When local minimum arrives in the training, the weights of NN become idle. If the chaotic variation of learning rate (LR) is included during training, the weight update may be accelerated in the local minimum...
A forecasting model for gas emission based on wavelet neural network is proposed in this paper. In the model, wavelet neutral network (WNN) is applied to the forecasting with gradient descent and amended by validity of iteration training algorithm. Compared with back-propagation neural networks, forecasting of the model has advantages of faster convergence and more accurate. Simulation results have...
In this paper we proposed a new algorithm for neural network training. This algorithm is developed from modification on Levenberg-Marquardt algorithm for MLP neural network learning. The proposed algorithm has good convergence. This method reduces the amount of oscillation in learning procedure. We named this algorithm as GK-LM Method. An example is given here to show usefulness of this method. Finally...
The feedforward neural networks trained with the online backpropagation (BP) learning algorithm have been widely studied in various areas of scientific research and engineering applications. In this paper we further study the convergence property of the online BP learning algorithm. Unlike the existing convergence analysis mainly focusing on the convergence of the gradient sequence of the error functions,...
This paper investigates the performance of conjugate gradient algorithms with sliding-window approach for training multilayer perceptron (MLP). Online learning is implemented when the system under investigation is time varying or when it is not convenient to obtain a full history of offline data about the system variables. Sliding window framework is proposed to combine the robustness of offline learning...
In this paper a new Lyapunov based backpropagation (BP) algorithm is proposed. The original BP algorithm is consists of one part, the learning rate factor (LR). In this new algorithm two extra adaptive parts has been added in comparison to original BP. The idea of adding these two parts is originated from the conventional PID controller. The first and second parts in the proposed algorithm are derivative...
In this paper, a faster supervised algorithm (BPfast) for the neural network training is proposed that maximizes the derivative of sigmoid activation function during back-propagation (BP) training. BP adjusts the weights of neural network with minimizing an error function. Due to the presence of derivative information in the weight update rule, BP goes to `premature saturation' that slows down the...
Through building up the functional relationship between the error E and the learning rate η, we propose one kind of new improved learning rate BP algorithm. This improved BP algorithm adopts serial dynamic adaptive learning rate, thus according to different error E to determine different learning rates. Compared with VLBP, the simulation result shows this improved variable learning rate BP algorithm...
Recently, neural network has attracted a great deal of attention as a method and theory realizing the artificial intelligence. According to the mechanism of BP neural network, the disadvantages of it are analyzed and a new improved BP algorithm is proposed in this paper. The improved algorithm is applied to the learning of XOR question and character recognition question. The simulations show that...
Face recognition is a front complex subject, which involves Physiology, Psychology, Image Processing, Computer Vision, Pattern Recognition and Mathematics. As a research success in the field of wavelet analysis theory, WNN(Wavelet Neural Network), a feed-forward network, avoids the blindness in structure design of BP(Back propagation) neural network, excludes the probability of sub-optimization in...
In order to overcome the disadvantages such as low calculation precision and convergence rate of traditional BP neural network algorithm, a kind of nonlinear optimization method-BFGS method for unconstrained extreme problem is introduced into BP neural network algorithm, and a BFGS-BP neural network model is developed, which is applied well in structure deformation monitoring data processing and forecasting...
In this paper, based on investigating and analyzing on the affecting factors for support type of development roadways as well as the successful support cases in Chengchao Iron mine, the improved BP neural network is put forward to study on the support type of development roadways. It may be seen from the learning course of learning samples and the prediction results of support types that whether the...
Image segmentation is critical to image processing and pattern recognition, An image segmentation system is proposed for the segmentation of color image based on neural networks. First, we introduce BP Neural network, it has the capacity of parallel computing, distributed saving, self-studying, fault-to-learnt and nonlinear function approximating. So it widely used in image segmentation, but it also...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.