Lecture 3

Bu lt- n matr x funct ons,

Read ng data from f les

Bu lt- n Matr x Funct ons

Matlab Funct on Descr pt on
 Determines the maximum value/values in x.
x is function argument (parameter).
max (x)


 If x is vector, the answer is a scalar value.

 If x is matrix, answer is vector conta n ng max mum of each column.
 Returns the largest value in vector x, to var able a.
[a b] = max (x)  Optionally b is the index locat on where the largest occurs.
 If x is matrix, returns largest value in each column, to vector a.

max (x, y)  Returns a matrix full of the larger of the two values in x and y.
m n (x)  Determines the smallest value in x.
sum (x)  Calculates the sum of elements in x.
prod (x)  Calculates the product (multiplication) of elements in x.
Returns a vector with values of x, in ascending order (default).
sort (x)


 If x is matrix, sort returns columns in ascending order.

sort (x, 'descend')  Sorts n decreasing order.
Returns the r-by-c matrix into B matrix, whose elements are taken
B = reshape (A, r, c)


column-wise from A matrix.

2
 Example : sum of vector
 When the argument of sum funct on s a vector, then the
result w ll be a scalar (s ngle value).

>> a = [10 20 30 40 50];

>> sum(a)

ans =
150

Example : sum of matr x

 When the argument of sum funct on s a matr x, then the result w ll be a vector.
 By default, the sum funct on works columnw se (totals of each column
separately).

>> a = [10 20 30;

40 50 60;
70 80 90] Overall Total of matr x
>> sum(sum(a))
ans =
>> sum(a) Totals of each 450
ans = column
120 150 180

>> sum(a, 2)
Totals of each row
ans =
(Parameter 2 means
60 roww se sum)
150
240 4
 Example : cumsum Funct on
 When argument s a vector, cumsum funct on returns a vector of
cumulat ve sums.
 When argument s a matr x, funct on returns a vector of
cumulat ve sums for each column.

>> a = [10 20 30 40 50]; >> a=[10 20 30;

40 50 60;
70 80 90]
>> cumsum(a)
ans =
10 30 60 100 150 >> cumsum(a)
ans =
10 20 30
50 70 90
>> sum(a) 120 150 180
ans =
150

Example : max Funct on

In examples below, X s a vector.
The max funct on returns only one scalar result by default.

>> X = [40 90 50 90 20]; We can d rectly g ve vector datas as

parameter.
>> max(X)
>> max( [40 90 50 90 20] )
ans =
ans =
90 90

>> [eb ind] = max(X)

 In follow ng example, eb s the max mum value, eb =
and nd s the ndex locat on of max mum. 90
 When there are more than one max mum (two 90’s),
Matlab returns the f rst element ndex. ind =
2

6
 Example : mean Funct on
Mean funct on returns the average of elements n a vector
(or averages n a matr x).

>> mean( [10 40 15] )

ans =
21.6667

Examples w th Matr x
c = [1 3 7 ;
2 8 4 ;
6 -1 -2];

>> min(c)
ans = Returns m n mum of each column
1 -1 -2

>> min(c, [], 2)

ans =
Returns m n mum of each row
1
(Parameter 2 means roww se operat on)
2
-2

>> mean(c)
ans = Returns average of each column
3.0000 3.3333 3.0000

8
 Example : Matr x Sort ng
 The default d mens on s columnw se, and the default order ng s the
ncreas ng order.
 Sort ng does not mod fy the or g nal matr x.
 It only generates a sorted new copy of the matr x.

>> c = [1 3 7 ;
2 8 4 ;
6 -1 -2];

Sorts each column

>> sort(c) separately
ans =
1 -1 -2
2 3 4
6 8 7

General syntax of Sort funct on

Y = sort (X, DIM, MODE)

X s vector or matr x that w ll be sorted.

DIM and MODE are optional parameters.
Y is sorted result, which has same d mens ons as X.

DIM selects a dimension along which to sort.

1 means columnw se sort ng (default)
2 means roww se sort ng.

MODE selects the direction of sort.

'ascend' results in ascending order (default).
'descend' results in descending order.

10
 Examples : Matr x sort ng
>> sort(c, 'descend')
>> c = [1 3 7 ; ans =
2 8 4 ; 6 8 7
6 -1 -2]; 2 3 4
1 -1 -2
Sorts each column separately n decreas ng order

>> sort(c, 2)
ans =
Sorts each row separately n ncreas ng order
1 3 7
(2 means roww se sort)
2 4 8
-2 -1 6

>> sort(c, 2, 'descend')

ans =
Sorts each row separately n
7 3 1 decreas ng order
8 4 2
6 -1 -2
11

Matr x row sort ng as groups : sortrows funct on


The sortrows funct on sorts rows of an matr x as groups n ascend ng order.

B = sortrows (A) sorts the rows of the matr x A n ascend ng order as a group.
B s the sorted result matr x w th the same s ze as A.

B = sortrows (A, COL) sorts A based on the columns spec f ed n the vector COL.
If an element of COL s pos t ve, the correspond ng column n A w ll be sorted n
ascend ng order.
f an element of COL s negat ve, the correspond ng column n A w ll be sorted n
descend ng order.
Example : sortrows (A, [2 -3] ) sorts the rows of A f rst n ascend ng order for the
second column, and then by descend ng order for the th rd column.

>> c = [1 3 7 ;
2 8 4 ;
6 -1 -2];

>> sortrows(c, [3])

ans = Sorts the rows as groups
6 -1 -2 of c matr x for the column 3,
2 8 4 n ascend ng order.
1 3 7
12
 Example : reshape Funct on
 The reshape funct on bu lds a new matr x.
 Number of elements n A and B are the same.
 Elements are collected columnw se.

>> A=[10 20 30 40 ;
50 60 70 80 ;
90 100 110 120]

>> B = reshape(A, 6, 2)
B =
10 30
50 70
90 110
20 40
60 80
100 120
13

Example : Calculat ng Sum of Taylor Ser es

 The following is formula of Taylor series , for calculating cos(x).


Wr te a Matlab program to read N (number of terms) from user,
then calculate Taylor ser es sum.

Test your program for X=150 degree.

Calculate the actual result (factual) by us ng the bu lt- n cos(X) funct on.

Calculate the truncat on error.

Absolute d fference between Taylor sum and the actual value s called the truncat on error.
Truncat on error = | factual ─ fsum |

14
 Matlab Program

% The program uses Taylor series to calculate cos(x) func on.

clc; clear;
N = input ('Terim sayisini (N) veriniz :');

degree = 150;
X = degree*pi/180;
% Degree is converted to Radian

k = [0 : N]; % Vector
terim = (-1).^k .* X.^(2*k) ./ factorial(2*k) ;
fsum = sum( terim )

factual = cos(X)
truncation_error = abs(factual – fsum)

15

Screen outputs

Test ng w th N=3 Test ng w th N=20

Terim sayisini (N) veriniz : 3 Terim sayisini (N) veriniz : 20

fsum = fsum =
-0.9168 -0.8660

factual = factual =
-0.8660 -0.8660

truncation_error = truncation_error =
0.0508 0

Number of terms (N) determ nes

the accuracy of result.

16
 Stat st cal Funct ons
Matlab
Descr pt on
Funct on
mean (x) Calculates the average of elements n vector x.

var (x) Calculates the var ance of x vector.

std (x) Calculates the standard dev at on of x vector.

mad (x) Calculates the mean absolute dev at on of x vector.

X  (x  x)
N N
2

Mean ( x )  1 Var ance  1

N N

Standard dev at on  Var ance

 x x
N

Mean absolute dev at on  1

N
17

Example: Stat st cal Funct ons

Veriler = [92 62 70 51];

Ortalama = mean (Veriler)

Varyans = var (Veriler)
Standart_dev = std (Veriler)
Mutlak_dev = mad (Veriler)

Ortalama =
68.7500

Varyans =
Screen 300.9167
output
Standart_dev =
17.3469

Mutlak_dev =
12.2500
18
 Random number generat ng funct on: rand
Matlab
Descr pt on
Funct on
 Generates nxn square matr x w th randomly
generated double (fract onal) data numbers.
rand (n)  Range of generated random numbers are
between 0.0 to 1.0

rand (n, m)  Generates nxm matr x w th random data.

>> rand (3) >> round ( rand (3) * 100 )

ans = ans =
0.3500 0.6160 0.8308 60 69 8
0.1966 0.4733 0.5853 26 75 23
0.2511 0.3517 0.5497 65 45 91

>> rand (1, 3) >> round ( rand (1, 3) * 100 )

ans = ans =
0.9172 0.2813 0.7572 92 28 76
19

Methods of
Read ng Data
from F les

20
 Methods of Read ng Data from F les

Method1: load funct on

 Read nto a matr x

Method2: textread funct on

• Read nto parallel vectors
• Alternat ve: Read nto a structure

21

Method1 : load Funct on

 When the data f le conta ns only numbers, the eas est method s the load funct on.

 EXAMPLE: Each row n f le conta ns data about a c rcle.

• x coord nate of center
• y coord nate of center
• rad us c rcles.txt F le

3 4 20
 The follow ng command reads all data into a 1 1 2
matrix, w th the same name of f le (c rcles). 7 5 10
>> load('circles.txt') 3 9 8
3 8 6
 The follow ng command reads data into a matrix
that named d fferently (da reler). 4 9 6
>> daireler = load('circles.txt') 1 6 3
5 5 12
4 4 15
22
 Access ng Matr x Columns

load ('circles.txt')
x = circles (: , 1) c rcles s a matr x, loaded from f le.
y = circles (: , 2) x, y, r are column vectors.
r = circles (: , 3)

Column 1 Column 2 Column 3

x = y = r =
3 4 20
1 1 2
7 5 10
3 9 8
3 8 6
4 9 6

23

Method2 : textread Funct on

• When data f le conta ns d fferent types of data n columns

(such as str ngs and numbers), the load funct on can not be used.
• Instead, the textread funct on should be used.
• Example: Read entire file into three column vectors.

>> [ogrnum, isim, notu] = textread ('students.txt', '%d %s %d' );

Vectors File name Format specifiers

d is decimal number
s is string
• First %d is for ogrnum
• Second %d is for notu
• %s is for isim

24
 Example Data F le : students.txt

Each row conta ns data about a student: 111 AAA 78

• ogrnum 222 BBB 85
• sm 333 CCC 89
• notu 444 DDD 94
555 EEE 86

ogrnum = sm= notu =

Workspace 78
111 'AAA'
(Three column 85
222 'BBB'
vectors) 333 'CCC' 89
444 'DDD' 94
86
555 'EEE'

25

Access ng tems from column vectors

Gett ng f elds from column vectors, for student at ndex 1.
>> ogrnum(1) >> isim(1) >> notu(1)

ans = ans = ans =

'AAA'
111 78

Gett ng f elds from column vectors, for student at ndex 2.

>> ogrnum(2) >> isim(2) >> notu(2)

ans = ans = ans =

'BBB'
222 85

26
 Import ng Data from
Spreadsheet F le
(M crosoft Excel, etc.)
nto MATLAB

27

Prepar ng CSV F le n Excel

• Open the Excel program and enter data values.
• From the F le menu, choose Save As.
• Select the CSV (Comma Separated Values) format.

28
 Import ng CSV F le nto Matlab
• In Matlab, select the Current Folder.
• R ght-cl ck on the source f le name (c rcles.csv).
• Cl ck the Import Data command.

c rcles.csv F le
x;y;r
3;4;20
1;1;2
7;5;10
3;9;8
3;8;6
4;9;6
1;6;3
5;5;12
4;4;15
29

Import ng CSV F le nto Matlab

• Matlab opens the Import w ndow.
• Select the mport ng cho ce (Column vectors or Matr x)
• Cl ck the Import Select on command.

30
 Import ng CSV F le nto Matlab
• After the mport ng, the Matlab Workspace conta ns the data values.
• Depend ng on the user select on, data w ll be n column vectors,
or n one matr x.

31