Support Vector

Machines
Here is where the presentation
begins

INDEX.HTML
“If you possess a restricted amount of information for
solving some problem, try to solve the problem directly and
never solve a more general one as an intermediate step. It is
possible that the available information is sufficient for a
direct solution but it is insufficient for solving a more
general intermediate problem.”

—SVM inventor, Vladimir VAPNIK

INDEX.HTML
<WHOA!
> Members:
Mehyar MLAWEH
Sami FARHAT
Ghada SOLTANI
Fide BEN SLIMEN
Youssef MOUSSA

INDEX.HTML
/Summary

/01 /Introduction /02 /Inputs

Historical? What is exactly What is the proper format for
SVM? input data?

/03 /Principles of Operation /04 /Use Case

How it works? The field use requirements

INDEX.HTML
/01
/
INTRODUCTI
ON
Let’s Discover the SVM

INDEX.HTML
/
INTRODUCTI
ON
SVM is a supervised learning Algorithm
developed by Vladimir Vapnik and it was first
heard in 1992

INDEX.HTML
Before we continue with the Svm, let’s first understand this term

/SUPERVISED
ALGORITHMS
In supervised learning, the training data provided to the
machines work as the supervisor that teaches the
machines to predict the output correctly. It applies the
same concept as a student learns in the supervision of
the teacher.

The aim of a supervised learning algorithm is to find a

mapping function to map the input variable(x) with the
output variable(y).

INDEX.HTML
/BASIC IDEA
The algorithm creates a line or a hyperplane
which separates the data into classes.

INDEX.HTML
/02
/INPUTS
What is the proper format for input data?

INDEX.HTML
/INPUTS & OUTPUTS

/INPUTS /OUTPUT /kernel

SVM is generally provided
with continuous variables that
will be divided to testing and
training data.
It returns predicted or Takes data as input and
transform it into the required
classified continuous
form.
variables.
Different SVM algorithms use
different types of kernel
functions.

INDEX.HTML
/03
/Principles of
Operation
HOW IT WORKS?

INDEX.HTML
/CASES?

/LINEAR PROBLEM
SVM is used for linearly separable
data, which means if a dataset can be
classified into two classes by using a
single straight line.
/NON-LINEAR
PROBLEM
The nonlinear problems are mapped
into a high-dimensional feature space,
in which the linear classification is
possible.

INDEX.HTML
/How does SVM work? (LINEAR PROBLEM)
/How does SVM work?
 /SVM’s way to find the best line

According to the SVM algorithm we find the points closest to the line from both the classes.
These points are called support vectors. Now, we compute the distance between the line and
the support vectors. This distance is called the margin. Our goal is to maximize the margin.
The hyperplane for which the margin is maximum is the optimal hyperplane.

INDEX.HTML
/How does SVM work?
/How does SVM work? (NON-LINEAR PROBLEM)
KERNEL Trick

Back to original
Kernel Functions Space
The idea is to transform the input data representation space into a higher dimensional space
using kernel Trick*, where a linear classifier can be used and achieve good performance.

Kernel Trick:

The knowledge of K(x,y ) makes it possible to calculate a scalar product where Φ(x) intervenes
without knowing the expression of Φ(y)

However, K(x,y) must satisfy certain conditions (Mercer conditions) for the corresponding Φ ( )
to exist.

INDEX.HTML
/ PROCESS

/Importing the /Splitting the /Training the SVM /Predicting the

dataset dataset into the Classification Test set results
Training set and model on the
Test set Training Set

/Visualizing the /Confusion Matrix

Results

INDEX.HTML
/04
/USE CASE
The field use requirements

INDEX.HTML
<SVM IS WIDELY
USED>
INDEX.HTML
/USES
/Face detection
SVMc classify parts of the image
as a face and non-face and create a
square boundary around the face
/Text and
/Classification
Hypertext
of images
Categorization
SVMs allow Text and hypertext
SVMs provides better search
categorization for both inductive
accuracy for image classification.
and transductive models.

/Generalized
/Handwriting /
predictive
recognition Bioinformatics
control(GPC)
We use SVMs to recognize It includes protein classification Use SVM based GPC to control
handwritten characters used widely. and cancer classification chaotic dynamics with useful
parameters.

INDEX.HTML
/THANKS!
/DO YOU HAVE ANY QUESTIONS?

Mehyar > mehyarmlaweh0@gmail.com

Samy > sami.farhat@ihec.ucar.tn
Ghada > ghada.soltani01@gmail.com
Fida > fidabenslimen3@gmail.com
Youssef > youssef.moussa@ihec.ucar.tn

INDEX.HTML