Evaluation
Evaluation
ANTRODUCTION
Model Evaluation is an essential part of the model development process. It helps to choose the
best model that represents our data and how well the chosen model will work in the future.
the stage which will help to decide that model is working well or not. In the evaluation process, o
evaluate the model by feeding the testing data and then comparingthe actual answer.
There can be different Evaluation techniques, depending of the type and purpose of the moe
important thing is that we cannot use the training data to test the model for evaluation. Tne
predict the
because the model will remember the whole training data and will therefore always
correct answer/output for any point in the training set. This process is known as over
1.2 MODEL EVALUATION TERMINOLOGY
In model evaluation, we come across many terms like:
machine.
1. Prediction / Predicted Value: The Prediction is the output which is given by the
2. Reality I ActualValue: The Actual Value represents the actual result of the Al Moue
Tounderstand the process of evaluation,we need to take an example for evaluation.
Case 1: We have trained the model for identifving an orange The image shown was ot an ofa9
Prediction matches with the
The model predicted Yes, which means the fruit is orange. The
Reality. Hence, this condition is termedas True Positive.
Case 2: We have given a lemon and asked the machine whether it is an orange of ot
Therefore. this condition is termed as
case, the machine too has predicted it correctly as a No.
True Negative.
model. But the machine has incorrectly
Case 3: Again, we displayed an image of a lemon to the
predicted that it is an orange. This case is termed as False Positive.
again we displayed an image of a orange and because of which the Reality is Yes
Case 4: Now is a
machine has incorrectly predicted it as a No, which means the machine predicts that it
but the
Negative.
lemon. Therefore, this case becomes False
1.3CONFUSION MATRIX
known as Error matrix and is represented by a table which describe
The confusion matrix is also
between the prediction and reality. The confusion matrixX allows us to
the result of comparison four parameters of the matrix.
understand the prediction results. Let us check the
POSITIVE (TP): Prediction matches the reality (True) and is actually positive.
TRUE
NEAANE(FP): Prediction does not match reality (False) and is actually negative.
FALSE
NEGATIVE (TN): Prediction matches the reality (True) and is actually negative.
TRUE
Prediction does not match reality (False) and is actually positive.
FALSE POSITIVE (FP):
[1]
f. The Emi NEGATIVE
|POSITIVE
FP
TP
POSITIVE FN TN
NEGA TIVE
Predicted
Values Positive (1) Negatve (0)
Positlve (1) TP FP
Negative (0) TN
In Percentage,
(TP +TN)
Number of correct predictions FP + FN)
x 100%
In Percentage, TP
Precision = TP+ FP x100%
In Percentage,
TP
Recall = X 100%
TP + FN
1.4.1Which motric is important?
The decision to choose precision or recall as an evaluation parameter completely depends upon
the conditions in which the model has been deployed. Procision is used as a metric when our
objective is to minimize falso positives and Recallis usod when the objoctive is to minimize
false negatives.
4. F1 Score: It is a measure of balance between precision and recall. It is computed as per
the following formula:
Precisionx Recal
F1 = 2 X
Precision + Recall
1.4 Comparesteitchnolwithothe
gy database
of Computer
Vision as it
of innages and gives the
result.
search
The COMPUTER
area of VISION TASKS see and
perceive.
Many
what istimes,
Computer humans
we look at Vision deals with objects and its surrounding9 that experienceto
an
determine
The
missing inoomplete
from it. All these task image and use our knowBedge
come under the category of
and
Computer Vision.
following are the main objects
1. Computer Vision tasks related to single and multiple
Classification: Image
belongs to. It is appliedclassification
only for singlebasicalty
objects.means
For exanmple. class the
object
identifyingthere are objects present
what
belonging to mnany classes such as ball, cat, sheep etc. The machine or deep learning
model would
2. deteminethat the animal detected in the image belongs to class cat.
Classification
identifying
+
Localisation:
and locating the object Classification processes of
present in that image. It is applied only for single objects
+ Localisation invoves both
3. Object Detection: It is the
along with their ability to detect or identfy objects in any given image correctly
the fom of three-dimensional
It is applied rectangular
position in the given Image. The 3D image is shown init.
boxes (known as Bounding Boxes) which bound the object within
onty for muitiple objects.
4. Instance Segmentation: Instance Segmentation basicaly
identiy
in the image with their
boundarnies intact. In this case, it's not onlydifferent instances g
about finding objects
an image, but also about creating a mask for each
detected obiect that is as accurate 2s
possible. It is applied only for multiple objects.
Classification Instanca
Classification Object Detection
+ Localization Segmentation
Mue s
Singie object
1.5 IMAGE BASICs images. Wel, you know that computers use
about how computers see the Digital Images thatyou see on
Computer vision is all images.
capture images, which are digital graphic) or by plotting some
to (bitmap
cameras as their eyes joining millions of small dots
created via
screens are either graphic).
mathematical equations (vector
1. Pixel The small dots
element) reptesents thethat create an image ale called pixels
terms of colour and smallest piece of the screen that be pixel (short for picture
(brightness). intensty. Each plxel can be sot to a controlled individually in
can
2. different colour and intensy
Resolution: The number of
measures is dimension of anpixelsoverall
the in an image
image ordetermine
its
resolution measUred in width how big the resolution
size
Image resolution
3. Pixol
Value: (vertical of an
The images stored in a pixels) and height (horizontal pixels).
image is Image
with numerical computer are
intensity of values, which are also called composed of pixels. Pixels
color and 255each pixel. They are integers that Pixel Values. These pixel values are denoted
4. represents white color. range from 0to 255, where 0represents represent the
Grayscale black
without Images: Grayscale images are images which have a
Color or apparent color, The darkest
zero value of pixel. The possible shade is black, whichrange of shades of gray
of color or 255 lightest is the total
value of a pixel, possible shade is white, which is the absence of
brightness levels of the three primaryIntermediate shades of gray are represented total presence
5. RGB
Images: colors. by equal
Colored
All other colors are images are made up of three primary
colors - red, green and blue.
images are stored generated by
in the form ofcombining different intensities- of these colors only. RGB
three different channels
of these channels red, green, and blue. Each
have
intensity of the pixels. values ranging from 0 to 255 where these numbers represent the
1. Fraud and Risk Detection: Fraud can be detected by using Data Science monitoring and
analysing of the user activity to find any unusual or malicious pattern. For instance,
abnormal variance in credit card usage or banking transactions can be easily trace so that
appropriate action maybe taken on time.
service provided
2. Internet Search or Autocomplete: Information retrieval is one of the best
Ask, etc., make use of
by search engines. Different search engines such as Google, Bing, of
the queries within a fraction
data science algorithms and give the best results for
seconds.
or
Website Recommendation: While doing online shopping on Amazon and Flipkart
3.
Netflix, you must have seen recommended products or videos that suit your taste.
watching
(5)
CLASS X
This is the
recommernder
activity of users and systern, which fnakas uss data s A r , arayzes tt
4.
Diagnosis recormmendsthe relsvart sst Aiterrs to the users.
of Modical
This can be done by Imaging:
plays an important role in Undor the dornain A data scienA, prosvArve a r a t s
prediction from it. analyzingirnproving the
the patient'health of patierts and rnanaging chronc diseaes
5. Digital s history. recoynizing the patterns and generatirg
digital Markoting and Target
marketing
customers. These
help
companies Advertising:
or The Dat Sciene related applications in
applications
related to their past make useorganizations
of addvertise their produts to the right
and alyorithrris
advertisements
1.3 DATA COLLECTION
search
accordingly. Thishistory
is also known as
that analyze the data of ustomers
interest. Aftef that, thhey urate and shm
For Data target advertisirvg
Science, usually the data is
stored in different
Some of the formats. collected in the form of tables These tabular
1. CSV: commonly used formats are:
datasets can te
CSV
tabular data.stands
Each
for Comnma
line of this fileSeparated
fields which are is a data Values. t is a
comma, hence theyseparated
are
by
cormnas. record and eachsirnple
Since record
file forrnat used to store
2. known as CSV the values of onsists of one or rnore
Spreadsheet:
in
as form Spreadsheet
tabular
such
is a
files.
using rows and computer
records are separated by a
Microsoft colurmns. application used for anatysis and
SQL: Structured Excel, LibreOffice
3. Calc,There
etc.
are many
storage of daa
soreadshet rograms
used for storing, Query Language is one of availabie
System (RDBMS)manipulating and
retrieving
such as MySql, SOL
the nost
the data inpopular languages for
UNIT -III Server, etc. Relational Database DeMS. It is
1.1 INTRODUCTION BASIC ICT SKILLS Maragement
An
Operating system is a
operating system is its program
(EMPLOYABI LITY SKILLS)
automatically the usermost
designed
important
to run other
1.2
when switch prograns on a computer. A
on the program Operating
FUNCTIONS OF THE computer. prograns on a cornputer'
system s
1. It
manages all the
OPERATING SYSTEM starts
whether it is busy or devices
not.
of a
computer and keeps track of the
2. It also
3. It also
checks whether the device is status of the device,
controls software functioning
resources the conputer.
of property or not.
ASS X
use by
4. It manages the computer memory and keeps track of which memory spaceis in
which program and which space is free.
5. It keeps track of the amount of disk space used by a specific file
6. It allows you to create, copy, move and delete files.
1.3TYPES OF OPERATING SYSTEMS
Broadly there are two types of operating system interfaces:
1. Command Line Interface (CLI)
[7)
ASS
X
4. Solarle: The Solaris Operating Systom, uBually known simply as Solaris, is a tree UnA
CLASS X
Mobilo Oporating Systom8: Those days mobile devices have turned smart, Gourtes,
6. operating systems are: Andro
smart mobile operating systoms. Some COmmon mobile
Apple iOS eto.
Windows 7is an
operating system developed by Microsoft and is used on personal computers
of Windows 7 Dosktop
1.5.1.1 Compononts
The components of Windows 7 Desktop are:
1. My Computor/ This PC: It displays all the storage areas of the computer.
2. Taskbar: Taskbar is the long horizontal bar present at the bottom of the screen.
3. Start button: It
is located on the left of the taskbar. When we Click the Start button, the
Stat menu opens and proVides access to programs and features.
A Recvcle Bin: Files and tolders deleted by the user are stored in Recycle Bin.
FOLDERS
1.6 FILES AND
The directory is also known as Tolder. The folder/directory structure is basically a storage location
or sub folders.
where we save our files
storane
Each directorv or folder can have files and sub-directory or sub-folders. A file is a named
Iocation where data is stored. This way of organizing files or folders is known as hierarchical file
that identifies the fle vne
structure. Each file is given a file name and has a file name extension
1.7 IMPORTANCE OF CARE AND
MAINTENANCE OF COMPUTER
Computer is an electronic device so it needs maintenance for its effective functioning. If the
maintenance is poor than there are chances of system failure.
Computer virus, malware, spyware, Trojan and other harmful software can ham the computer
system and lead to failure of hardware. If antivirus is installed in the computer, and is updated and
run periodically, then we can save our data from any kind of loss.
1.8BASIC TIPS FOR TAKING CARE OF DEVICES
Computer components needs proper care so that the devices stay safe for a longer duration.
Some of the precautions to be taken while cleaning the computer components are:
(8)
Never spray or
the liquid onto a squirt any
cotton cloth.liquid onto any computer
component. If a spray is neede0. spra
When cleaning a
component of the computer, turn it off before
When cleaning, be careful to not cleaning.
accidentally adjust any knobs or controis
UNIT- GREEN SKILLS
V
1.1 SUSTAINABLE DEVELOPMENT (EMPLOYABILITY SKILLS)
SUstainable development is a development that meets the needs of the present without
Compromising the ability of future generations to meet their own needs.
our dally lite we can contribute to create a Sustainable Society by following 4Rs' and 1U of
Sustainability.
Inese are REFUSE, REDUCE, REUSE, RECYCLE and UPCYCLE.
1. Refuse: We must first REFUSE to use products that may harm the environment. This wil
help in creating sustainable environment.
2. Reduce: REDUCE comes after REFUSE. In this step, we minimize the use of the products
that may cause harm to environment.
3. Reuse: Next Step is REUSE. In reusing we can use that resources in another way and
keeps them out of the trash.
we must try to
4. Recycle: Next step after REUSE is RECYCLE. After reusing the product,
recycle it, so that the leftover resource is used properly again.
5. Upcycle: UPCYCLING of products could be done manually asdesirable.well as with the help of
machines, giving a new look to the old product and making it look
DEVELOPMENT
1.2 IMPORTANCE OF SUSTAINABLE
resources
1. To keep a check on the over exploitation of natural
2. To regenerate the renewable energy resources
3. To ensure safety for human beings
4. To stop environmental degradation
5. Tofind alternative sources of energy