Shulou(Shulou.com)06/01 Report--

Editor to share with you what are the general functions of the neural network in matlab. I hope you will get something after reading this article. Let's discuss it together.

General function of Neural Network I. Neural Network Simulation function

[Ymai Pfrecinct Afrecy Ereco Perf] = sim (net,P,Pi,Ai,T)

Where parameter Y is the output of the network; Pf represents the final input delay state; Af represents the final layer delay state; E is the error between the actual output and the target vector; perf is the performance value of the network; net is the network object to be tested; P is the input vector matrix of the network; Pi is the initial input state delay state (which can be omitted); Ai is the initial layer delay state (which can be omitted) T is the target vector (which can be omitted).

2. Neural network training function

1.train function

[net,tr,Y,E,Pf,Af] = train (net,P,T,Pi,Ai)

The output parameter net is the trained network; tr is the training record; Y is the network output vector; E is the error vector; Pf is the input delay state at the end of training; Af is the layer delay state at the end of training. The input parameter net is the pre-training network, P is the input vector matrix of the network, T represents the target matrix of the network, the default value is 0 position Pi indicates the initial input delay, the default value is 0% Ai indicates the initial layer delay, and the default value is 0.

The network training function is a general learning function, which repeatedly applies a set of input vectors to a network and updates the network every time, knowing that a certain criterion has been achieved. the stop criterion may be to achieve the maximum learning steps, the minimum error gradient or the error goal.

2.trainb function

Used to train the weights and thresholds of the neural network. The calling format of the function is as follows:

Net.trainFcn = 'trainb'

This function cannot be called directly, but is implicitly called through the function train. Train calls trainb to train the network by setting the network property net.trainFcn to trainb.

P = [0 1 2 3 4 5]

T = [0 000 1 1 1]

Net = feedforwardnet (3)

Net = train (net,p,t)

Y=net (p)

3.adapt function

[net,Y,E,Pf,Af] = adapt (net,P,T,Pi,Ai)

The input parameter net is the neural network to be adapted; P is the network input; T is the network target, the default value is 0 position Pi is the initial input delay, and the default value is 0; the default value is 0; the output net parameter is the adaptive neural network, Y is the network output; E is the network error; Pf is the final input delay; Af is the final layer delay.

P = [- 1 0 1 0 11-1 0-1 01]

T = [- 1-1 1 1 20-1-1 0 11]

Net=linearlayer ([0 1], 0.1)

Create a linear layer with an input range of [- 11] and a learning rate of 0.1

[net,t,e,pf] = adapt (net,p,t)

Mse (e)

Neural network learning function 1.learnp function

[dW,LS] = learnp (W, P, P, Z, N, A, A, T, T, E, E, W, G, W, G, A, D, D, L, P, L, and L)

Parameter dW is the weight change matrix; LS is the current learning state; W is the weight matrix of SxR (may be omitted); P is the input vector matrix of RxQ; Z is the weight matrix of the input layer of SxQ (may be omitted); N is the network input matrix of SxQ (may be omitted); E is the error matrix (E=T-Y); T represents the objective vector of the network (may be omitted); A represents the actual output vector of the network (which may be omitted). GW is the performance-related weight gradient matrix of SxR (may be omitted), gA is the performance-related output gradient matrix of SxQ (may be omitted), D is the neuron distance matrix of SxS (may be omitted), LP is the learning parameter (may be omitted), and LS learning function declaration (may be omitted).

Use learnp function to learn a perceptron network to make it complete the'OR 'function.

2.learn function

This function is also a weight and threshold learning function, but its virtual seat speed is much faster than that of learnp blocks when the input vector and copy change greatly or there are singular values.

Neural network initialization function 1.revert function

This function is used to restore the updated weights and thresholds to the last initialized values. The calling format of the function is net = revert (net).

If the network result has changed, that is, if the link relationship between the weight and threshold of the network, as well as the input and the length of each layer are different from the original network structure, then the function cannot restore the weights and thresholds to the original values.

2.init function

Net = init (net)

3.initlay function

Net = initlay (net) is used to initialize a layer of a linear neural network

4.initnw function

Net = initnw (net, I) this function is a layer initialization function, and I is a hierarchical index

Neural network input function 1.netsum function

This function is an input summation function, which adds the weighted input and threshold of a layer as the input of that layer, and the function is called in the following format:

N = netsum ({Z _ 1 ~ Z _ 2 ·J ~ Zn}, Fp)

The parameter Z1, Fp, Z2, Zn represents the first input, its number can be arbitrary, and the parameter Z1 is an array of functional parameter units, which is ignored in the course.

2.netprod function

The calculation of this function is similar to that of netsum, except that it is a quadrature function that multiplies the weight and threshold of a layer as input to that layer.

N = netprod ({Z _ 1, Z _ 2, Z _ 2, pen, Zn})

3.concur function

The function of this function is to make the structure of the weight vector and the threshold vector consistent, so that the addition or multiplication can be carried out. The calling format of the function is as follows:

Concur (BBJQ)

Where B is the weight vector of Nx1 dimension and Q is the length needed to achieve homogenization.

VI. Transfer function of neural network

1.hardlim function

A = hardlim (NMagi FP)

When the input vector matrix N of the network is given, the output vector matrix An of this layer is returned. When the element in N is greater than or equal to 00:00, the return value is 1, otherwise it is 0.

2.hardlims function

The output values are-1 and 1

After reading this article, I believe you have a certain understanding of "what are the general functions of neural networks in matlab". If you want to know more about it, you are welcome to follow the industry information channel. Thank you for reading!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.