Qing Hui Wu

A MATLAB-Based Study on Approximation Performances of Improved Algorithms of Typical BP Neural Networks

BP neural networks are widely used and the algorithms are various. This paper studies the advantages and disadvantages of improved algorithms of five typical BP networks, based on artificial neural network theories. First, the learning processes of improved algorithms of the five typical BP networks are elaborated on mathematically. Then a specific network is designed on the platform of MATLAB 7.0 to conduct approximation test for a given nonlinear function. At last, a comparison is made between the training speeds and memory consumption of the five BP networks. The simulation results indicate that for small scaled and medium scaled networks, LM optimization algorithm has the best approximation ability, followed by Quasi-Newton algorithm, conjugate gradient method, resilient BP algorithm, adaptive learning rate algorithm. Keywords: BP neural network; Improved algorithm; Function approximation; MATLAB

A Study on Approximation Performances of Improved Bp Neural Networks Based on LM Algorithms

When approximating nonlinear functions, standard BP algorithms and traditional improved BP algorithms have low convergence rate and tend to be stuck in local minimums. In this paper, standard BP algorithm is improved by numerical optimization algorithm. Firstly, the principle of Levenberg-Marquardt algorithm is introduced. Secondly, to test its approximation performance, LMBP neural network is programmed via MATLAB7.0 taking specific nonlinear function as an example. Thirdly, its approximation result is compared with those of standard BP algorithm and adaptive learning rate algorithm. Simulation results indicate that compared with standard BP algorithm and adaptive learning rate algorithm, LMBP algorithm overcomes deficiencies ranging from poor convergence ability, prolonged convergence time, increasing iteration steps to nonconvergence. Thus with its good approximation ability, LMBP algorithm is the most suitable for medium-sized networks.

Application of Probabilistic Neural Network in Pattern Classification

The network model of probabilistic neural network and its method of pattern classification and discrimination are first introduced in this paper. Then probabilistic neural network and three usually used back propagation neural networks are established through MATLAB7.0. The pattern classification of dots on a two-dimensional plane is taken as an example. Probabilistic neural network and improved back propagation neural networks are used to classify these dots respectively. Their classification results are compared with each other. The simulation results show that compared with back propagation neural networks, probabilistic neural network has simpler learning rules, faster training speed and it needs fewer training samples; the pattern classification method based on probabilistic neural network is very effective, and it is superior to the one based on back propagation neural networks in classifying speed, accuracy as well as generalization ability.

Approximation Performance of BP Neural Networks Improved by Heuristic Approach

Among all improved BP neural network algorithms, the one improved by heuristic approach is studied in this paper. Firstly, three types of improved heuristic algorithms of BP neural network are programmed in the environment of MATLAB7.0. Then network training and simulation test are conducted taking a nonlinear function as an example. The approximation performances of BP neural networks improved by different numerical optimization approaches are compared to aid the selection of proper numerical optimization approach.

Pattern Classification Based on Self-Organizing Feature Mapping Neural Network

Traditional pattern classification methods are not always efficient because sample data sets are sometimes incomplete and there are exceptions and counter examples. In this paper, SOFM neural network is applied in pattern classification of two-dimensional vectors after analysis of its structure and algorithm. The method to establish SOFM network via MATLAB7.0 is introduced before the network is applied to classify two-dimensional vectors. The adjustment process of weight vectors together with classification performance of SOFM model are also tested in the condition of different number of training steps. The simulation results show that the classification approach based on SOFM model is effective because of its fast speed, high accuracy and strong generalization ability.

A Study on Approximation Performances of General Regression Neural Network

In order to study the approximation performance of general regression neural networks, the structure and algorithm of general regression neural networks are first introduced. Then general regression neural networks and back propagation neural networks improved by Levenberg-Marquardt algorithm are established through programming using MATLAB language. A certain nonlinear function is taken as an example to be approximated by the two kinds of neural networks. The simulation results indicate that compared with back propagation neural networks, general regression neural networks has better approximation precision and faster convergence speed, which means it has much better approximation ability than back propagation neural networks. Therefore, for more complex function approximation, general regression neural networks is recommended. It can reduce the complexity of neural networks and it is also easier to design.

Application of Probabilistic Neural Networks in Fault Diagnosis of Three-Phase Induction Motors

In fault diagnosis of three-phase induction motors, traditional methods usually fail because of the complex system of three-phase induction motors. Short circuit is a very common stator fault in all the faults of three-phase induction motors. Probabilistic neural network is a kind of artificial neural network which is widely used due to its fast training and simple structure. In this paper, the diagnosis method based on probabilistic neural network is proposed to deal with stator short circuits. First, the principle and structure of probabilistic neural network is studied in this paper. Second, the method of fault setting and fault feature extraction of three-phase induction motors is proposed on the basis of the fault diagnosis of stator short circuits. Then the establishment of the diagnosis model based on probabilistic neural network is illustrated with details. At last, training and simulation tests are done for the model. And simulation results show that this method is very practical with its high accuracy and fast speed.

A Study on the Application of Learning Vector Quantization Neural Network in Pattern Classification

Standard back propagation (BP) neural network has disadvantages such as slow convergence speed, local minimum and difficulty in definition of network structure. In this paper, an learning vector quantization (LVQ) neural network classifier is established, then it is applied in pattern classification of two-dimensional vectors on a plane. To test its classification ability, the classification results of LVQ neural network and BP neural network are compared with each other. The simulation result shows that compared with classification method based on BP neural network, the one based on LVQ neural network has a shorter learning time. Besides, its requirements for learning samples and the number of competing layers are also lower. Therefore it is an effective classification method which is powerful in classification of two-dimensional vectors on a plane.

Sensor Fault Diagnosis Based on SOFM Neural Network

Traditional sensor fault diagnosis is mainly based on statistical classification methods. The discriminant functions in these methods are extremely complex, and typical samples of reference modes are not easy to get, therefore it is difficult to meet the actual requirements of a project. In view of the deficiencies of conventional sensor fault diagnosis technologies, a fault diagnosis method based on self-organizing feature map (SOFM) neural network is presented in this paper. And it is applied to the fault diagnosis of pipeline flow sensor in a dynamic system. The simulation results show that the fault diagnosis method based on SOFM neural network has a fast speed, high accuracy and strong generalization ability, which verifies the practicality and effectiveness of the proposed method.

Application of General Regression Neural Network in Characteristic Curve Fitting of Optical Fiber Micro-Bend Sensor

In order to reflect the input and output features of an optical fiber micro-bend sensor, a new method using general regression neural network (GRNN) to fit the characteristic curve is proposed in this paper. First, the measuring principle of optical fiber micro-bend sensor and the principle of GRNN are introduced. Then, to verify the feasibility and effectiveness of this new method, a comparison between two kinds of fitting methods is done. One is based on GRNN, the other is based on Levenberg-Marquart improved BPNN. The results of the simulation experiment show that with the same number of training samples and for small scale to medium scale networks, compared with BPNN, GRNN has smaller error, faster convergence speed and higher fitting accuracy. So the method discussed in this paper provides a reliable basis for the nonlinear compensation problem of optical fiber micro-bend sensor.