Introduction to Neural Network

toolbox in Matlab

Matlab stands for MATrix LABoratory.

Matlab 5.3.1 with toolboxs.
• SIMULINK
• Signal Processing Toolbox
• Control System Toolbox
• System Identification Toolbox
• Robust Control Toolbox
• Spline Toolbox
• Optimization Toolbox
• Neural Network Toolbox
• Image Processing Toolbox
• Symbolic Math Toolbox
• Fuzzy Logic Toolbox
• Statistics Toolbox
• Wavelet Toolbox
• Communication Toolbox
• Database Toolbox
 Programming Language : Matlab

 High-level script language with interpreter.

 Huge library of function and scripts.
 Act as an computing environment that combines
numeric computation, advanced graphics and
visualization.
 Entrance of matlab

Type matlab in unix command prompt

• e.g. sparc76.cs.cuhk.hk:/uac/gds/username> matlab

• If you will find an command prompt ‘>>’ and you have successfully entered matlab.
>>
Ask more information about
software

>> info
– contacting the company
• eg. Technique support, bugs.
>> ver
– version of matlab and its toolboxes
– licence number
>> whatsnew
– what’s new of the version
 Function for programmer

 help : Detail of function provided.

– >> help nnet, help sumsqr
 lookfor : Find out a function by giving some keyword.
– >> lookfor sum
•TRACE Sum of diagonal elements.
•CUMSUM Cumulative sum of elements.
•SUM Sum of elements.
•SUMMER Shades of green and yellow colormap.
•UIRESUME Resume execution of blocked M-file.
•UIWAIT Block execution and wait for resume.
……………...
Function for programmer (cont’d)

which : the location of function in the system

(similar to whereis in unix shell)
– >> which sum So that you can save it
– sum is a built-in function. in your own directory and
– >> which sumsqr modify it.
– /opt1/matlab-5.3.1/toolbox/nnet/nnet/sumsqr.m
Function for programmer (cont’d)

! : calling unix command in matlab system

– >> !ls
– >> !netscape
 Plotting graph

 Visualisation of the data and result.

 Most important when handing in the report.
 plot : plot the vector in 2D or 3D
– >> y = [1 2 3 4]; figure(1); plot(power(y,2));
Add vector x as the
x = [2 4 6 8]; x-axis index
Index of the vector (you
plot(x,power(y,2));
can make another vector
for the x-axis)
 Implementation of Neural
Network using NN Toolbox
Version 3.0.1

 1. Loading data source.

 2. Selecting attributes required.
 3. Decide training, validation, and testing data.
 4. Data manipulations and Target generation.
– (for supervised learning)
 5. Neural Network creation (selection of network architecture) and initialisation.
 6. Network Training and Testing.
 7. Performance evaluation.
 Loading data

load: retrieve data from disk.Save variables in matlab

environment and load back
– In ascii or .mat format.
>> data = load(‘wtest.txt’);
>> whos data;
Name Size Bytes Class
data 826x7 46256 double array
 Matrix manipulation

for all

Start for 1
stockname = data(:,1);

1 2
2 4
training = data([1:100],:)
a=[1;2]; a*a’ => [1,2;2,4];
a=[1,2;2,4]; a.*a => [1,4;4,16];
1 4
4 16
 Neural Network Creation and
Initialisation

 net = newff(PR,[S1 S2...SNl],{TF1 TF2...TFNl},BTF,BLF,PF) S2: number of

 Description ouput neuron
 NEWFF(PR,[S1 S2...SNl],{TF1 TF2...TFNl},BTF,BLF,PF) takes,
 PR - Rx2 matrix of min and max values for R input elements. S1: number
 Si - Size of ith layer, for Nl layers. hidden neurons
 TFi - Transfer function of ith layer, default = 'tansig'.
 BTF - Backprop network training function, default = 'trainlm'.
 BLF - Backprop weight/bias learning function, default = 'learngdm'.
 PF - Performance function, default = 'mse’ and returns an
 N layer feed-forward backprop network.
-1 1 neuron 1
>> PR = [-1 1; -1 1; -1 1; -1 1]; -1 1 Number of inputs
Min
-1 1 Max decided by PR
-1 1
 Neural Network Creation

 newff : create a feed-forward network. TF2: logsig

 Description
 NEWFF(PR,[S1 S2...SNl],{TF1 TF2...TFNl},BTF,BLF,PF) takes,
 PR - Rx2 matrix of min and max values for R input elements. TF1: logsig
 Si - Size of ith layer, for Nl layers.
 TFi - Transfer function of ith layer, default = 'tansig'.
 BTF - Backprop network training function, default = 'trainlm'.
 BLF - Backprop weight/bias learning function, default = 'learngdm'.
 PF - Performance function, default = 'mse’ and returns an
 N layer feed-forward backprop network.

>> net = newff([-1 1; -1 1; -1 1; -1 1], [4,1], {‘logsig’ ‘logsig’}); Number of inputs

decided by PR
 Network Initialisation

Initialise the net’s weighting and biases

>> net = init(net); % init is called after newff

re-initialise with other function:

– net.layers{1}.initFcn = 'initwb';
– net.inputWeights{1,1}.initFcn = 'rands';
– net.biases{1,1}.initFcn = 'rands';
– net.biases{2,1}.initFcn = 'rands';
 Network Training

 The overall architecture of your neural network is store in the

variable net;
 We can reset the variable inside.
net.trainParam.epochs =1000; (Max no. of epochs to train) [100]
net.trainParam.goal =0.01; (stop training if the error goal hit) [0]
net.trainParam.lr =0.001; (learning rate, not default trainlm) [0.01]
net.trainParam.show =1; (no. epochs between showing error) [25]
net.trainParam.time =1000; (Max time to train in sec) [inf]
 Network Training(cont’d)

 train : train the network with its architecture.

 Description -0.5 1 -0.5 1
-1 0.5 -1 0.5
 TRAIN(NET,P,T,Pi,Ai) takes,
0.5 1 0.5 1
For
 NET - Network. Training -0.5 -1 -0.5 -1 neuron 1
 P - Network inputs. pattern 1
 T - Network targets, default = zeros.
 Pi - Initial input delay conditions, default = zeros.
 Ai - Initial layer delay conditions, default = zeros.

>> p = [-0.5 1 -0.5 1; -1 0.5 -1 0.5; 0.5 1 0.5 1; -0.5 -1 -0.5 -1];
 Network Training(cont’d)

 train : train the network with its architecture.

 Description -1 1 -1 1
 TRAIN(NET,P,T,Pi,Ai) takes,
 NET - Network.
Training
 P - Network inputs.
 T - Network targets, default = zeros.
pattern 1
 Pi - Initial input delay conditions, default = zeros.
 Ai - Initial layer delay conditions, default = zeros.

>> p = [-0.5 1 -0.5 1; -1 0.5 -1 0.5; 0.5 1 0.5 1; -0.5 -1 -0.5 -1];

>> t = [-1 1 -1 1];

>> net = train(net, p, t);

 Simulation of the network

 [Y] = SIM(model, UT)

 Y : Returned output in matrix or structure format.

 model : Name of a block diagram model.
 UT : For table inputs, the input to the model is interpolated.

-0.5 1.00
-0.25 1.00
-1.00 0.25
For
Training -1.00 0.50 neuron 1
pattern 1

>> UT = [-0.5 1 ; -0.25 1; -1 0.25 ; -1 0.5];

>> Y = sim(net,UT);
 Performance Evaluation

 Comparison between target and network’s output in testing set.(generalisation ability)

 Comparison between target and network’s output in training set. (memorisation

ability)

 Design a function to measure the distance/similarity of the target and output, or

simply use mse for example.
 Write them in a file
(Adding a new function)

Create a file as fname.m (extension as .m)

>> fname
function [Y , Z] = othername(str) loading.m

Y = load(str);
Z = length(Y);

>> [A,B] = loading('wtest.txt');

 Reference

Neural Networks Toolbox User's Guide

– http://www.cse.cuhk.edu.hk/corner/tech/doc/manual/matlab-5.3.1/help/pdf_doc/nnet/nnet.pdf

Matlab Help Desk

– http://www.cse.cuhk.edu.hk/corner/tech/doc/manual/matlab-5.3.1/help/helpdesk.html

Mathworks ower of Matlab

– http://www.mathworks.com/