Module 1:

Introduction to
Machine Learning
Agenda
1. How is Machine Learning different from Data Mining
2. Key Tasks of Machine Learning
3. Types of Machine Learning
4. Issues in Machine Learning
How is Machine
Learning different
from Data Mining?
How is Machine Learning different
from Data Mining?
Data mining
 It is a subset of business analytics and refers to
exploring an existing large dataset to unearth
previously unknown patterns, relationships and
anomalies that are present in the data.
 It gives us the ability to find completely new insights
that we weren’t necessarily looking for.
 For example, if a business has a lot of data on
customer churn, it could apply a data mining
algorithm to find unknown patterns in the data and
identify new associations that could indicate
customer churn in the future. In this way, data
mining is frequently used in retail to spot patterns
and trends.
How is Machine Learning different
from Data Mining?
Machine learning
 Machine learning is a subset of artificial
intelligence (AI).
 With machine learning, computers analyze large
data sets and then ‘learn’ patterns that will help it
make predictions about new data sets.
 Apart from the initial programming and maybe
some fine-tuning, the computer doesn’t need
human interaction to learn from the data.
 Netflix predicting you may want to watch Ozark
next, based on the viewing preferences of other
users with similar profiles, is an example of
machine learning in action.
 Real-time fraud detection on credit card
transactions is another example.
Types of Machine
Learning
What are different Machine Learning Models ?
Categories of ML algorithms:

 Supervised Machine Learning

 Learning with labelled data.

 Unsupervised Machine Learning

 Learning with unlabelled data.

 Reinforcement Learning
 Learning through reward
maximization.

Wolfewicz, A. (2020, November 26). A beginner’s guide to how machines learn - Levity. Medium. Retrieved November 7, 2021, from https://medium.com/levity/a-beginners-guide-to-how-machines-learn-bd161bad3328

7
Supervised Learning Model

8
Supervised Learning

 Supervised Learning indicates having a supervisor as a

teacher

 E.g. A math teacher teaches the class by making the

students learn using a lot of solved examples(training) and
then tests the knowledge gained by giving the class,
problems to solve on their own

 Similarly, Supervised Learning is when you make the

machine learn by teaching or training the machine using
labeled data

Download Teach Clipart Teaching Style, Picture - Teacher Clipart - Png Download (#5747301) - PinClipart. (n.d.-b). PinClipart.Com. Retrieved May 2022, from https://www.pinclipart.com/downpngs/ibmoTxx_teach-clipart-teaching-style-picture-teacher-clipart-png/

9
Supervised Learning- Example

 Let’s consider the example of currency coins

 Problem Statement: Build a model to predict the coin based on
its weight Feature Label
 Assume that we have different currency coins (dataset) having
different weights
 1 Euro weighs 5 grams Weights Currency
5 1 Euro
 1 Dirham weighs 7 grams
 1 Dollar weighs 3 grams 7 1 Dirham
 1 Rupee weighs 4 grams and so on 3 1 Dollar
4 1 Rupee
 Feature – Weights
 Label – Currency

10
 3 Unseen
Feature Label grams coin

Test
Weights Currency
5 1 Euro Train
7 1 Dirham Supervised Learning
Model
3 1 Dollar
4 1 Rupee
Prediction

1 Dollar
Supervised Learning- Example

13
Unsupervised Learning Model

14
Unsupervised Learning

 Unsupervised Learning is a type of learning without any

guidance

 For e.g. A child learning to swim on his own without any

supervision. Here, the child is the model trying to discover
ways and techniques to swim and the swimming pool is
similar to the unknown data fed to the model.

 Here, the machine is responsible to discover patterns,

similarities, and differences on its own based on the
unlabeled dataset.
Students PNG Images | Vector and PSD Files | Free Download on Pngtree. (n.d.). Retrieved May 2022, from Pngtree website: https://pngtree.com/so/students

15
Unsupervised Learning - Example
Let’s consider the example of a supermarket
Assume that we have a customer database with records of their products bought over
a period.
Now you being the marketing manager decides to send a grocery offer message to that
customer who buys grocery regularly.
Customer ID Total number Total number
of grocery of non-grocery
items bought items bought Grocery
shopper
Unsupervised
Learning Model
Non-
grocery
shopper
 Note that there was no customer labeled as a grocery shopper or non-grocery shopper.

 Model could discover patterns on its own and could produce these two clusters/groups.
16
Unsupervised Learning

(n.d.-b). Retrieved May 2022, from Btechmag.com website: https://btechmag.com/wp-content/uploads/2020/11/UNSupervised-Learning.gif

18
Let’s have a quick knowledge
check!

19
Identify the model: Supervised or
Unsupervised ?
Case 1: Facebook identifies your friend in a picture from an album of
tagged photographs

Supervised OR Unsupervised

An Update About Face Recognition on Facebook - About Facebook. (2019, September 3). Retrieved from About Facebook website: https://about.fb.com/news/2019/09/update-face-recognition/

20
Identify the model: Supervised or
Unsupervised ?
Case 1: Facebook identifies your friend in a picture from an album of
tagged photographs

Supervised OR Unsupervised

An Update About Face Recognition on Facebook - About Facebook. (2019, September 3). Retrieved from About Facebook website: https://about.fb.com/news/2019/09/update-face-recognition/

21
Identify the model: Supervised or
Unsupervised ?
Case 2: Netflix Recommendations based on someone’s watch
history

Supervised OR Unsupervised

Netflix makes finding new shows and films easier with “play something” recommendation feature. (n.d.). Retrieved May 2022, from The Financial Express website: https://www.financialexpress.com/industry/technology/netflix-makes-finding-new-shows-and-films-easier-with-

play-something-recommendation-feature/2241987/
22
Identify the model: Supervised or
Unsupervised ?
Case 2: Netflix Recommendations based on someone’s watch
history

Supervised OR Unsupervised

Netflix makes finding new shows and films easier with “play something” recommendation feature. (n.d.). Retrieved May 2022, from The Financial Express website: https://www.financialexpress.com/industry/technology/netflix-makes-finding-new-shows-and-films-easier-with-

play-something-recommendation-feature/2241987/
23
Identify the model: Supervised or
Unsupervised ?
Case 3: Analyze bank data for suspicious-looking transactions and flag the fraud
transactions
[Note that suspicious transactions are not defined in this case]

Supervised OR Unsupervised

SBI, ICICI, other banks see hike in suspicious transactions: Report. (2020, September 11). Retrieved from Deccan Herald website: https://www.deccanherald.com/business/business-news/sbi-icici-other-banks-see-rise-in-suspicious-transactions-report-885

24
Identify the model: Supervised or
Unsupervised ?
Case 3: Analyze bank data for suspicious-looking transactions and flag the fraud
transactions
[Note that suspicious transactions are not defined in this case]

Supervised OR Unsupervised

SBI, ICICI, other banks see hike in suspicious transactions: Report. (2020, September 11). Retrieved from Deccan Herald website: https://www.deccanherald.com/business/business-news/sbi-icici-other-banks-see-rise-in-suspicious-transactions-report-8

25
Issues in Machine
Learning
Issues in Machine Learning

1. Inadequate Training Data:

 Noisy Data, Incorrect data and

Generalizing of output

2. Poor quality of data

 Noisy data, incomplete data, inaccurate

data, and unclean data lead to less accuracy
in classification and low-quality results.
Hence, data quality can also be considered
as a major common problem while
processing machine learning algorithms.

27
Issues in Machine Learning

3. Non-representative training data

 To make sure our training model is generalized well or not, we

have to ensure that sample training data must be
representative of new cases that we need to generalize. The
training data must cover all cases that are already occurred as
well as occurring.

 Further, if we are using non-representative training data in the

model, it results in less accurate predictions.

28
Issues in Machine Learning
4. Overfitting and Underfitting (discuss later)
5. Monitoring and maintenance
 regular monitoring and maintenance become
compulsory for the same. Different results
for different actions require data change;
hence editing of codes as well as resources
for monitoring them also become necessary.
6. Getting bad recommendations
 A machine learning model operates under a
specific context . If the context changes it
results in bad recommendations and leads to
concept drift in the model.
29
Issues in Machine Learning

7. Lack of skilled resources

 The absence of skilled resources in the form of

manpower is also an issue. Hence, we need manpower
having in-depth knowledge of mathematics, science,
and technologies for developing and managing scientific
substances for machine learning.

30