Papers by Keyword: BP Algorithm

Abstract: BP algorithm is a classical neural network algorithm. We analyzed the deficiency of traditional BP neural network algorithm, designed new S function and momentum method strategy, optimized the algorithm parameters. We use the new algorithm in the classification of orange images, take color and shape features as input value, the experimental results proved that our algorithm is faster and the classification accuracy rate reaches to 90%

Abstract: Artificial neural network has unique advantages for massively parallel processing, distributed storage capacity and self-learning ability. The paper mainly constructs neural network identifier and neural network controller for system identification and control on temperature and humidity of heating and drying system of materials. And the paper introduces the structure and principles of neural network, and focuses on analyzing learning algorithm, training algorithm and limitation of the most widely applied multi-layer feed-forward neural network (BP network), based on which the paper proposes introducing momentum to improve BP network.

Abstract: The advantages and weakens of traditional BP algorithm is briefly analyzed and an efficient global optimization algorithm is proposed.The basic principle of the algorithm is presented,and a new BP neural network algorithm based on the existing BP algorithm and the new global optimization algorithm is proposed, considering the new global optimization algorithm can solve the problem of local minimum efficiently. To verify the effectiveness of the new BP algorithm,the paper compared the experimental results of various algorithms in solving function fitting problem.

Abstract: Parallelized training algorithm of MLP-BP neural network is implemented on the Sunway Blue Light Supercomputer. Efforts are mainly focused on the dada parallel method based on the characteristics of the training process. The implementation mainly depends on MPI techniques, which ensures the universality of the application. A new strategy of data partition and storage is applied in the parallelism. Tests of short-term traffic prediction which is significant in the intelligent transportation system are carried out to verify the accuracy and efficiency of the routines.

Abstract: The problem of local minimum cannot be avoided when it comes to nonlinear optimization in the learning algorithm of neural network parameters, and the larger the optimization space is, the more obvious the problem becomes. This paper proposes a new type of hybrid learning algorithm for three-layered feed-forward neural networks. This algorithm is based on three-layered feed-forward neural networks with output layer function, namely linear function, combining a quasi Newton algorithm with adaptive decoupled step and momentum (QNADSM) and iterative least square method to export. Simulation proves that this hybrid algorithm has strong self-adaptive capability, small calculation amount and fast convergence speed. It is an effective engineer practical algorithm.

Abstract: Due to the instability and low precision of electromagnetic position trackers and the inefficiency of existing calibrating methods, a method with high accuracy and effectiveness for FOB (Flock of Birds) calibration was studied. The components, operational principle, merits and drawbacks of FOB were briefly introduced. The positions of 343 sampling points set in the effective working area were measured and the data was processed for trainings and tests of the calibration model established using genetic algorithm and BP algorithm. Experiments were conducted to verify the effectiveness of the method and the results showed the calibrated tracker’s average errors in the X, Y, and Z direction were 0.86cm, 0.70cm and 0.83cm respectively, meeting the requirements of human-computer interaction.

Abstract: In order to reduce the appearance of aviation material mishap, it is important to predict the aviation material mishap for safety management and decision-making body. Considering the advantage of neural network modeling, an aviation material mishap prediction based on neural network and its BP algorithm model is proposed. An actual example on fight mishap 10000-Hour-Rate data of USAF illustrates that the proposed prediction model has an accurate prediction.

Abstract: It is an important work for modern libraries to predict reader flow. With the help of reader flow, library staff can grasp the change regulation of readers, allocate tasks rationally and take steps ahead of time in high-risk period. Because of reader flows typical non-linear characteristics, evolutionary neural network technology is introduced in this research so as to improve the accuracy of reader flow prediction. A prediction method for library reader flow based on evolutionary neural network is proposed. Genetic algorithm is used to optimize and design BP neural network firstly, then evolutionary neural network is used to predict reader flow. The experimental results show that evolutionary neural network is an effective tool for us to predict library reader flow. We can realize an accurate prediction for library reader flow by this method.

Abstract: This paper puts emphasis on studying on artificial neural network (ANN) method. Following that a model of wind power prediction is established based on back-propagation neural network. In order to improve the learning speed of ANN, a revised BP algorithm is adopted by using variable step and the combination of GA and BP algorithm. This method has a good effect in practice.

Abstract: In this paper, the bipartite graphical structure of LDPC codes is studied, some methodology for designing shorten LDPC Codes are presented. The performance evaluation of shorten LDPC codes presented provides easy and effective strategy in selecting appropriate shorten scheme, avoiding the complex encoding and iterative decoding steps. The shorten LDPC codes whose minimum code weight greater and length 4-cycle free have better performance.