can u pls explain the logic of this program in detail?

clc;
clear all;
close all;
warning off all;

%%Load the normal data

load data1
load data2
load data3
load data4
load data5
load data1_1
load data1_2
load data1_3
load data1_4
load data1_5
load data2_1
load data2_2
load data2_3
load data2_4
load data2_5
load data3_1
load data3_2
load data3_3
load data3_4
load data3_5

T = [data1,data2,data3,data4,data5,data1_1,data1_2,data1_3,data1_4,data1_5,data2_1,data2_2,data2_3,data2_4,data2_5,data3_1,data3_2,data3_3,data3_4,data3_5];
x = [0,0,0,0,0,1,1,1,1,1,2,2,2,2,2,3,3,3,3,3];

%%create a feed forward neural network

net1 = newff(minmax(T),[30 20 1],{'logsig','logsig','purelin'},'trainrp');
net1.trainParam.show = 25;
net1.trainParam.lr = 0.01;
net1.trainParam.epochs = 3000;
net1.trainParam.goal = 0;

%%train the neural network using the input,target and the created network

[net1] = train(net1,T,x);

%%save the network

save net1 net1

%%simulate the network for a particular input

y = round(sim(net1,T))

Answer 1

[Greg Heath]

This is a very poorly designed classifier.

My advice is to ignore it and study the classification examples in the help and doc documentation of your MATLAB version

 help patternnet % or the obsolete newpr
 doc patternnet

Also search the NEWSGROUP and ANSWERS for posted code. For example, search using

 greg patternet 
 greg newpr
 greg classify
 greg classifier
 greg classification

If you find code that interests you, post your interpretation and I will check it for you.

Greg

Answer 2

[Walter Roberson]

Not a chance. It would take several pages just to describe in detail what the "clear all" does. The code is commented. If you don't know what feed forward networks are, then research feed forward networks. If you do not understand MATLAB constructs such as net1.trainParam.lr then study MATLAB instead of asking us to write a detailed explanation.

http://www.mathworks.com/matlabcentral/answers/8026-best-way-s-to-master-matlab

[Jan]

@ASHA: Please take into account that Walter is very good in explaining. When he decides to suggest the "best ways to master matlab" thread, this is a really good answer.