Location:
Search - tansig
Search list
Description: 神经网络训练根据Kolmogorov定理，输入层有14个节点，所以中间层有29个节点
％中间层神经元的传递函数为 tansig
％输出层有8个节点，其神经元传递函数为logsig
％训练函数采用traingdx-neural network training under the Kolmogorov theorem, input layer has 14 nodes, Therefore, the intermediate layer has 29% of nodes middle layer neurons in the transfer function for the output layer tansig% have eight nodes, its neuron transfer function for the training function logsig% used traingdx
Platform: |
Size: 873 |
Author: 陈胜 |
Hits:
Description: 利用MATLAB对神经网络进行编程,用newff()创建两层前向网络。网络输入范围[-1 1]，第一层有10个tansig神经元-using MATLAB right neural network programming with newff () to the creation of a two-tier network. Network input range [-1 1], the first layer 10 tansig neurons
Platform: |
Size: 4941 |
Author: 龙海侠 |
Hits:
Description: 利用MATLAB对神经网络进行编程,用newff()创建两层前向网络。网络输入范围[-1 1]，第一层有10个tansig神经元-using MATLAB right neural network programming with newff () to the creation of a two-tier network. Network input range [-1 1], the first layer 10 tansig neurons
Platform: |
Size: 5120 |
Author: 龙海侠 |
Hits:
Description: 神经网络训练根据Kolmogorov定理，输入层有14个节点，所以中间层有29个节点
％中间层神经元的传递函数为 tansig
％输出层有8个节点，其神经元传递函数为logsig
％训练函数采用traingdx-neural network training under the Kolmogorov theorem, input layer has 14 nodes, Therefore, the intermediate layer has 29% of nodes middle layer neurons in the transfer function for the output layer tansig% have eight nodes, its neuron transfer function for the training function logsig% used traingdx
Platform: |
Size: 1024 |
Author: 陈胜 |
Hits:
Description: 倒立摆是一种复杂、时变、非线性、强耦合、自然不稳定的高阶系统，许多抽象的控制理论概念都可以通过倒立摆实验直观的表现出来。基于人工神经网络BP算法的倒立摆小车实验仿真训练模型，其倒立摆BP网络为4输入3层结构。输入层分别为小车的位移和速度、摆杆偏离铅垂线的角度和角速度。隐含层单元数16个。输出层设置为1个输出单元。输入层采用Tansig函数，隐含层采用Logsig函数，输出层采用Purelin函数。用Matlab 6.5数值计算软件对模型进行学习训练，并与线性反馈控制逻辑算法对比，表明倒立摆控制BP算法精度高、收敛快，在非线性控制、鲁棒控制等领域具有良好的应用前景。 -Inverted pendulum is a complex, time-varying, nonlinear, strong coupling, the natural instability of the high-end systems, many of the abstract concept of control theory to pass through the inverted pendulum experiment demonstrated intuitive. Based on artificial neural network BP algorithm inverted pendulum experiment simulation training model car, the Inverted Pendulum BP network input 3-layer structure of 4. Input layer, respectively, for the car s displacement and speed of deviation from the plumb line placed under the angle and angular velocity. Hidden layer unit number 16. Output layer is set to an output unit. Tansig function using input layer, hidden layer Logsig function used, the output layer Purelin function. Numerical calculation using Matlab 6.5 software for learning and training model, and linear feedback control logic algorithm comparison, show that the inverted pendulum control of BP algorithm and high precision, fast convergence in nonlinear control, robust control and
Platform: |
Size: 217088 |
Author: 月到风来AA |
Hits:
Description: load training.txt load TrainOut.txt
load validation.txt load ValOut.txt
load testing.txt load TestOut.txt
INPUT=[training validation testing]
OUTPUT=[TrainOut ValOut TestOut]
net=newff(INPUT,OUTPUT,200,{ tansig , purelin }, trainlm )
Platform: |
Size: 1024 |
Author: 鄭又豪 |
Hits:
Description: this matlab code for estimating the static linear system(system function is time variable) with Recursive Least Squre and 2 solutions for better result.
1- using the Covariance Matrix Reseting in a specefic time.
2-using the RLS with Forget Factor
this program is written by matlab 7.0.
Here we want to estimate the below function:
1-u^2+(1+tansig(0.1*(t-375)))*u^3+u^5+3*u^7
finally,there are plots for showing results.
-this is matlab code for estimating the static linear system(system function is time variable) with Recursive Least Squre and 2 solutions for better result.
1- using the Covariance Matrix Reseting in a specefic time.
2-using the RLS with Forget Factor
this program is written by matlab 7.0.
Here we want to estimate the below function:
1-u^2+(1+tansig(0.1*(t-375)))*u^3+u^5+3*u^7
finally,there are plots for showing results.
Platform: |
Size: 2048 |
Author: maysam |
Hits:
Description: this matlab code for estimating the static linear system(system function is time variable) with Kalman Filter.
this program is written by matlab 7.0.
Here we want to estimate the below function:
this is matlab code for estimating the static linear system(system function is time variable) with Recursive Least Squre and 2 solutions for better result.
1- using the Covariance Matrix Reseting in a specefic time.
2-using the RLS with Forget Factor
this program is written by matlab 7.0.
Here we want to estimate the below function:
1-u^2+(1+tansig(0.1*(t-375)))*u^3+u^5+3*u^7
finally,there are plots for showing results.-this is matlab code for estimating the static linear system(system function is time variable) with Kalman Filter.
this program is written by matlab 7.0.
Here we want to estimate the below function:
this is matlab code for estimating the static linear system(system function is time variable) with Recursive Least Squre and 2 solutions for better result.
1- using the Covariance Matrix Reseting in a specefic time.
2-using the RLS with Forget Factor
this program is written by matlab 7.0.
Here we want to estimate the below function:
1-u^2+(1+tansig(0.1*(t-375)))*u^3+u^5+3*u^7
finally,there are plots for showing results.
Platform: |
Size: 1024 |
Author: maysam |
Hits:
Description: C++实现神经网络，包括三个函数，sigmoid，tanh，tansig的实现方式。-designed by C++，implication neural network，include sigmoid，tanh，tansig
Platform: |
Size: 2653184 |
Author: Air |
Hits:
Description: 神经网络的C++源程序。提供sigmod，tanh，tansig三种函数的实现方式。-Neural Networks C++ source code. Provide sigmod, tanh, tansig three functions are implemented.
Platform: |
Size: 335872 |
Author: Air |
Hits:
Description: This a simple program to calculate the output of artificial neural network (ANN) using microcontroller ATMega 32.
Assume that the neural architecture is : 2 hidden layers with 4 and 2 neurons
respectively and 1 layer output with 1 neuron.This program explains the step how to compute the ANN in off-line mode.Means
the weights and biases are already exist (microcontroller is not doing the learning steps). These parameters are produced using MATLAB. Therefore, we have already made the appropriate architecture using MATLAB. The tansig function is use in both of hidden layer.
Finally, the neural output will be displayed on LCD port C.-This is a simple program to calculate the output of artificial neural network (ANN) using microcontroller ATMega 32.
Assume that the neural architecture is : 2 hidden layers with 4 and 2 neurons
respectively and 1 layer output with 1 neuron.This program explains the step how to compute the ANN in off-line mode.Means
the weights and biases are already exist (microcontroller is not doing the learning steps). These parameters are produced using MATLAB. Therefore, we have already made the appropriate architecture using MATLAB. The tansig function is use in both of hidden layer.
Finally, the neural output will be displayed on LCD port C.
Platform: |
Size: 1024 |
Author: koeskoes |
Hits:
Description: 函数逼近的MATLAB程序，本程序设计一个两层的bp网络用于函数逼近，隐层的激活函数为 tansig,输出层激活函数为purelin线性函数
-Function approximation of the MATLAB program, the program design of a two-tier network for function approximation bp, hidden layer activation function is tansig, the output layer activation function is a linear function purelin
Platform: |
Size: 3072 |
Author: 冯玉玺 |
Hits:
Description: 三层BP神经网络对药品的销售进行预测。：输入层有四个结点，隐含层结点数为5，隐含层的激活函数为tansig；输出层结点数为1个，输出层的激活函数为logsig，并利用此网络对药品的销售量进行预测，预测方法采用滚动预测方式，即用前四个月的销售量来预测第四个月的销售量。-Three-layer BP neural network to forecast the sales of drugs. : Input layer has four nodes, 5 hidden layer nodes, hidden layer activation function tansig output layer nodes is 1, the output layer activation function is logsig, and use the network of drug sales volume to predict the prediction method using rolling forecasts, that is used to predict sales of the first four months the first four months of sales.
Platform: |
Size: 26624 |
Author: 张洁 |
Hits:
Description: T=[1 0 0 1 0 0 1 0 0
0 1 0 0 1 0 0 1 0
0 0 1 0 0 1 0 0 1]
输入向量的最大值和最小值
threshold=[0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1]
net=newff(threshold,[31 3],{ tansig , logsig }, trainlm )
训练次数为1000，训练目标为0.01，学习速率为0.1
net.trainParam.epochs=1000
net.trainParam.goal=0.01
LP.lr=0.1
net = train(net,P,T)
测试数据，和训练数据不一致
P_test=[0.2101 0.0950 0.1298 0.1359 0.2601 0.1001 0.0753 0.0890 0.0389 0.1451 0.0128 0.1590 0.2452 0.0512 0.1319
0.2593 0.1800 0.0711 0.2801 0.1501 0.1298 0.1001 0.1891 0.2531 0.0875 0.0058 0.1803 0.0992 0.0802 0.1002 -T = [1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 1 0 0 1] ' of the maximum and minimum input vector threshold = [0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1] net = newff (threshold, [31 3], {' tansig' , ' logsig' }, ' trainlm' ) training times for the 1000 target of 0.01 training, learning rate of 0.1 net.trainParam.epochs = 1000 net. trainParam.goal = 0.01 LP.lr = 0.1 net = train (net, P, T) test data, and training data inconsistencies P_test = [0.2101 0.0950 0.1298 0.1359 0.2601 0.1001 0.0753 0.0890 0.0389 0.1451 0.0128 0.1590 0.2452 0.0512 0.1319 0.2593 0.1800 0.0711 0.2801 0.1501 0.1298 0.1001 0.1891 0.2531 0.0875 0.0058 0.1803 0.0992 0.0802 0.1002
Platform: |
Size: 1024 |
Author: 王飞 |
Hits:
Description: function exp85
clear all
p=[0:0.1:1.1]
t=[22.4570 26.6012 12.6416 5.9367 6.9265 28.2432 31.5068 37.0166 7.8947 1.0398 12.7095]
net=newff([0 1],[5 1],{ tansig purelin }, traingdx , learngdm )
net.trainParam.epochs=2500
net.trainParam.goal=0.001
net.trainParam.show=50
net=train(net,p,t)
r=sim(net,p)
plot(p,t,p,r, * )
y=sim(net,[1.2])
-function exp85
clear all
p=[0:0.1:1.1]
t=[22.4570 26.6012 12.6416 5.9367 6.9265 28.2432 31.5068 37.0166 7.8947 1.0398 12.7095]
net=newff([0 1],[5 1],{ tansig purelin }, traingdx , learngdm )
net.trainParam.epochs=2500
net.trainParam.goal=0.001
net.trainParam.show=50
net=train(net,p,t)
r=sim(net,p)
plot(p,t,p,r, * )
y=sim(net,[1.2])
Platform: |
Size: 2048 |
Author: reynard |
Hits:
Description: 单输出函数Y=SIN(X)逼近问题的bp程序:假设网络结构为3--2--1，输入维数M,共N个样本,一般输入不算层，输出算层- 激活函数: hardlim---(0,1),hardlims---（-1,1),purelin,logsig---(0,1),tansig----(-1,1)
softmax,poslin,radbas,satlin,satlins,tribas
训练算法: 1.traingd,traingdm,traingda(variable learning rate
backpropagation),trainrp( resilient backpropagation )
2.conjugate gradient (traincgf, traincgp, traincgb, trainscg),
quasi-Newton (trainbfg, trainoss), and Levenberg-Marquardt (trainlm).
Platform: |
Size: 2048 |
Author: 刘老师 |
Hits:
Description: 神经网络预测程序 求一用 matlab 编的程序 P=[。。。] 输入 T=[。。。] 输出 创建一个新的前向神经网络 net_1=newff(minmax(P),[10,1],{ tansig -BP neural network prediction program with MATLAB
Platform: |
Size: 238592 |
Author: 李明 |
Hits:
Description: 人脸识别过程中的奇异值分解算法代码，亲测可用，实现步骤为：
feature = allFeature(1) //featurenumber=8,16,24,32,48,64,80
[pn,pnewn,t,num_train,num_test] = train_test(feature,num_train) //num_train=1~10
[net] = createBP(pn) //110,tansig,purelin,trainrp,1e-5,8000,0.005
[net,tr] = trainBP(net,pn,t)
[result_test,result_train,count_test,count_train,Test_reg,Train_reg,Total_reg] = result(net,pnewn,pn,num_train,num_test) -Recognition process singular value decomposition algorithm code, pro-test can be used to achieve the steps of: feature = allFeature (1) //featurenumber = 8,16,24,32,48,64,80 [pn, pnewn, t , num_train, num_test] = train_test (feature, num_train) //num_train = 1 ~ 10 [net] = createBP (pn) // 110, tansig, purelin, trainrp, 1e-5, 8000,0.005 [net, tr ] = trainBP (net, pn, t) [result_test, result_train, count_test, count_train, Test_reg, Train_reg, Total_reg] = result (net, pnewn, pn, num_train, num_test)
Platform: |
Size: 8075264 |
Author: 陈伟 |
Hits: