MATLAB

Programming I
Outline
Examples
What is Matlab?
History
Characteristics
Organization of Matlab Toolboxes
Basic Applications
Programming

Example I
Obtain the result of matrix multiply






Matlab Code
>>A=[4,6; 9,15; 8,7];
>>B=[3,24,5; 7,19,2];
>>C=A*B
C =
54 210 32
132 501 75
73 325 54
4 6
3 24 5
9 15
7 19 2
8 7
(
(
(

(
(

(

C Code
#include <cstdlib>
#include <iostream>
#include <fstream>
#include <sstream>
#include <iomanip>
#include <vector>

using namespace std;
typedef vector<vector<int> > Mat;
void input(istream& in, Mat& a);
Mat matrix_product(const Mat& a, const Mat& b);
void print(const Mat& a);

C Codes
int main(int argc, char *argv[])
{
ifstream in1 ( "Matrix A.txt" );
ifstream in2 ( "Matrix B.txt" );
int row, row1, col, col1;
in1>>row>>col;
in2>>row1>>col1;

Mat a(row, vector<int>(col));
input(in1, a);
Mat b(row1, vector<int>(col1));
input(in2, b);

print(matrix_product(a,b));
system("PAUSE");
return EXIT_SUCCESS;
}

C Code Cond.
void input(istream& in, Mat& a)
{
for(int i = 0; i < a.size(); ++i)
for(int j = 0; j < a[0].size(); ++j)
in>>a[i][j];
}

void print(const Mat& a)
{
for(int i = 0; i < a.size(); ++i)
{ cout<<endl;
for(int j = 0; j < a[0].size(); ++j)
cout<<setw(5)<<a[i][j];
}
cout<<endl;
}


C Code Cond.
Mat matrix_product(const Mat& a, const Mat& b)
{
Mat c(a.size(), vector<int>(b[0].size()));
for(int i = 0; i < c.size(); ++i)
for(int j = 0; j < c[0].size(); ++j)
{
c[i][j] = 0;
for(int k = 0; k < b.size() ; ++k)
c[i][j] = c[i][j] + a[i][k] * b[k][j];
}
return c;
}

Example II
Calculate the transpose of the matrix:




Matlab Code
>>A=[3,17,2;5,19,7;6,4,54]
>>A^(-1)
ans =
-0.9327 0.8505 -0.0757
0.2131 -0.1402 0.0103
0.0879 -0.0841 0.0262

3 17 2
5 19 7
6 4 54
(
(
(
(

What is MATLAB?
Abbreviation of MATrix LABoratory

Is a high-performance language for
technical computing

It integrates computation, visualization,
and programming in an easy-to-use
environment where problems and
solutions are expressed in familiar
mathematical notation.
Usage of MATLAB
Math and computation
Algorithm development
Data acquisition
Modeling, simulation, and prototyping
Data analysis, exploration, and visualization
Scientific and engineering graphics
Application development
including graphical user interface building


History
In 1970s Dr.Cleve Moler
Based on Fortran on EISPACK and LINPACK
Only solve basic matrix calculation and graphic
1984 MathWorks Company by Cleve Moler and
Jack Little
Matlab 1.0 based on C
Included image manipulations, multimedia etc.
toolboxes
In 1990 run on Windows
Simulink, Notebook, Symbol Operation, etc.
21 Century

Main Features
Powerful tool for matrix calculations
Basic data type is double complex matrix with subscript is 1
High efficient/reliable algorithms
Multi-paradigm
- interactive environment (no compiler)
- programming
Most input/output parameters in Matlab
functions are variable
User friendly with extensively help system

MATLAB SYSTEM
Develop Tools
Toolboxes
Data-Accessing Tools
Stateflow
Generating-Codes Tools
Module Libraries
Third parties modules
Applications
DATA
C
Simulink
MATLAB
Buildup of MATLAB
Development Environment
MATLAB desktop and Command Window, a command history, an editor
and debugger, and browsers for viewing help, the workspace, files, and the
search path.
The MATLAB Mathematical Function Library
A vast collection of computational algorithms ranging from elementary
functions to more sophisticated functions
The MATLAB Language
A high-level matrix/array language with control flow statements,
functions, data structures, input/output, and object-oriented
programming features
Graphics
Display vectors and matrices as graphs
two-dimensional and three-dimensional data visualization
image processing
animation and presentation graphics
The MATLAB Application Program Interface (API)
Interchange the C/Fortran codes with Matlab codes
Basic Applications I
Matrices and Arrays
Generating Arrays/Matrices
Elementary Matrices:
sum, transpose, diagonal, subscripts
Variables: does not require any type
declarations or dimension statements
Operators: +,-,*,/, ^, etc.
Examples of Expressions
Generating Matrices
Directly enter
A = [16 3 2 13; 5 10 11 8; 9 6 7 12; 4 15 14 1]
Use Functions




Example: B = zeros(1,5) row vector/array
C = ones(7,1) column vector/array
D = rand(5,8,3) multiple dimension matrix
Load/Import from files

zeros All 0s
ones All 1s
rand Uniformly distributed random elements
randn Normally distributed random elements
Elementary Matrices
>>D=rand(3,2,3)
D(:,:,1) = D(:,:,2) = D(:,:,3) =
0.9501 0.4860 0.4565 0.4447 0.9218 0.4057
0.2311 0.8913 0.0185 0.6154 0.7382 0.9355
0.6068 0.7621 0.8214 0.7919 0.1763 0.9169
>> sum(D(:,1,1))
ans =
1.7881
>> E=D(:,:,2)'
E =
0.4565 0.0185 0.8214
0.4447 0.6154 0.7919
>> a=diag(D(:,:,3))
a =
0.9218
0.9355
>> F=[D(1,:,1),D(2,:,3); D(1,:,1)+1, D(2,:,3)-1]
F =
0.9501 0.4860 0.7382 0.9355
1.9501 1.4860 -0.2618 -0.0645
>>F(:,2)=[]
F =
0.9501 0.7382 0.9355
1.9501 -0.2618 -0.0645
>> F(2:1:4)=[]
F =
0.9501 0.9355 -0.0645
Examples of Expression
>>rho = (1+sqrt(5))/2
rho =
1.6180

>>a = abs(3+4i)
a =
5

>>huge = exp(log(realmax))
huge =
1.7977e+308 =1.8X10
308
Basic Application II
Graphic

Programming
Flow Control
Selector: if, switch & case
Loop: for, while, continue, break, try & catch
Other: return
Other Data Structures
Multidimensional arrays, Cell arrays, Structures,
etc.
Scripts and Functions

Flow Control Examples I
Selector:
2-way:
if rand < 0.3 %Condition
flag = 0; %then_expression_1
else
flag = 1; %then_expression_2
end %endif
Multiple-ways:
if A > B | switch (grade)
'greater | case 0
elseif A < B | Excellent
'less | case 1
elseif A == B | Good
'equal | case 2
else | Average
error('Unexpected situation') | otherwise
end | Failed
| end

Flow Control Examples II
Loop
>>Generate a matrix
for i = 1:2
for j = 1:6
H(i,j) = 1/(i+j);
end
end
>>H
H =
0.5000 0.3333 0.2500 0.2000 0.1667 0.1429
0.3333 0.2500 0.2000 0.1667 0.1429 0.1250


Loop Cond.
>>%Obtain the result is a root of the polynomial x
3
- 2x 5
a = 0; fa = -Inf;
b = 3; fb = Inf;
while b-a > eps*b
x = (a+b)/2;
fx = x^3-2*x-5;
if sign(fx) == sign(fa)
a = x; fa = fx;
else
b = x; fb = fx;
end
end
>>x
x =
2.09455148154233

Scripts
MATLAB simply executes the commands found
in the file
Example
% Investigate the rank of magic squares
r = zeros(1,32); %generation a row array 1X 32
% ; tell the system does not display
% the result on the screen
for n = 3:32
r(n) = rank(magic(n));
end
r
bar(r) %display r in bar figure
Scripts Cond.
Note:
No input/output parameters
(Different from functions! )
Write the script in editor, which can be
activated by command edit in command
window, save it with its specific name
Next time you want to execute the script just
call the name in the command window


Functions
Functions are M-files that can accept input
arguments and return output arguments.
The names of the M-file and of the
function should be the same.
Function Example I
Background of AIC
AIC is used to compare the quality of nested
models
AIC requires a bias-adjustment in small sample
sizes


1
2 ln( ) 2
2 ( 1)
2 ln( ) 2
1
min( )
exp( 0.5 )
exp( 0.5 )
i i
i
i R
r
r
AIC likelihood K
K K
AICad likelihood K
n K
delta AIC AIC
delta
w
delta
=
= +
+
= + +

=

=

Function Head (Comments)

%AICad Function
%Date:Dec.29th, 2008
%Purpose: Calculate the adjusted
% Akaikes Information Criteria (AIC)
%InPut: AIC
% n (the sample size)
% K (the number of parameters in the model)
%OutPut: AICad
% delta
% weight
%Functions Called: None
%Called by: None

IN Command Window
>>AIC=[-123; -241.7; -92.4; -256.9];
>>n=12;
>>K=[4;8;2;6];
>>[AICad, delta, w] = CalAIC(AIC, n, K)
AICad = delta = w =

-83.0000 89.9000 0.0000
-97.7000 75.2000 0.0000
-80.4000 92.5000 0.0000
-172.9000 0 1.0000

M file
function[AICad,delta,w]=CalAIC(AIC,n,K)
AICad = AIC+2.*K.*(K+1);
a = min(AICad);
delta = AICad - a;
b = exp(-0.5.*delta);
c = sum(b);
w = b./c;
Function Example II
Background of PCA
Comparing more than two communities or samples

Sample Post Oak Red Oak Water Oak
1 66 25 24
2 48 30 31
3 26 55 42
4 15 20 49
5 11 50 58
combine post post red red water water
Y w x w x w x = + +
Background of PCA
Cond.
Principal Components Analysis (PCA)
It finds weighted sums of samples or
observations (linear combinations) that are
highly correlated with the values of the response
variables (e.g., species abundances)

PCA objectively chooses these weights in such a way
as to maximize the variance in the weighted sums

Based on correlations between response variables

.
Function Head
%Principle Component Analysis Function
%Date:Dec.29th, 2008
%Purpose: Obtain the axis scores of PCA
%InPut: the abundance of 3 species
% in 5 different samples
%OutPut: sample scores of PCA axis
%Functions Called: None
%Called by: None
IN Command Window
>>a=[66;48;26;15;11];
>>b=[25;30;55;20;50];
>>c=[24;31;42;49;58];
>>[score1, score2]=PCA(a,b,c)
score1 = score2 =
-2.2233 -0.1963
-1.1519 -0.0743
0.8734 -1.0888
0.4871 1.5120
2.0147 -0.1526


M file
function[x11,x12,x13,x21,x22,x23]=PCA(a,b,c)
%Calculate the mean and population standard deviation of 3 species
for i=1:5
d(i)=(a(i)-mean(a))^2;
stda=sqrt(sum(d)/5);
e(i)=(b(i)-mean(b))^2;
stdb=sqrt(sum(e)/5);
f(i)=(c(i)-mean(c))^2;
stdc=sqrt(sum(f)/5);
end
%Obtain the z-scores of 3 species
for i=1:5
za(i)=(a(i)-mean(a))/stda;
zb(i)=(b(i)-mean(b))/stdb;
zc(i)=(c(i)-mean(c))/stdc;
end
N=[za',zb',zc'];
M file Cond.

%Obtain the original correlation matrix
M=[1,corr(a,b),corr(a,c);corr(a,b),1,corr(b,c);corr(a,c),corr(b,c),1];
[V1,D1]= eig(M); %Calculate the eigenvector and eigenvalues
weight1=V1(:,3); %Get the maximum eigenvector
score1=N*weight1;

%We use the exact same procedure that we used for axis 1,
%except that we use a partial correlation matrix
%The partial correlation matrix is made up of all pairwise correlations
%among the 3 species, independent of their correlation with the axis 1
%sample scores. Hence, axis 2 sample scores will be uncorrelated
%with (or orthogonal to) axis 1 scores.
M file Cond.
new=[a,b,c,score1];
M2=[1,corr(a,b),corr(a,c),corr(a,score1);
corr(a,b),1,corr(b,c),corr(b,score1);
corr(a,c), corr(b, c), 1, corr(c,score1);
corr(a,score1),corr(b,score1),corr(c,score1),1];
for i=1:3
for j=1:3
corrnew(i,j)=M2(i,j);
end
end
g=[M2(4,1)*M2(4,1),M2(4,1)*M2(4,2),M2(4,1)*M2(4,3);
M2(4,2)*M2(4,1),M2(4,2)*M2(4,2),M2(4,2)*M2(4,3);
M2(4,3)*M2(4,1),M2(4,3)*M2(4,2),M2(4,3)*M2(4,3)];
corrmatrix=corrnew-g;
[V2,D2]=eig(corrmatrix);
weight2=V2(:,3); %signs needs to be checked with JMP
score2=N*weight2;
MATLAB in MCSR
Matlab is available on willow and sweetgum
www.mcsr.olemiss.edu
[email protected]

Enter matlab can activate Matlab in willow
or sweetgum
Use ver to obtain the version of Matlab,
which can tell you which toolboxes of
Matlab in that machine

Price of personal purchase is 500$ include update
releases (only include MATLAB & Simulink)
You need pay separate fee for each toolbox such as
statistical, bioinformatics, etc

Take Home Message