Supervised Learning
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Agenda
1 Why do we need Machine Learning?
2 What is Machine Learning?
3 Life cycle to build a model with ML
4 What is supervised Learning?
5 Naïve Bayes Algorithm
6
Logistic Regression
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
Why do we need Machine Learning?
• In the past, we used to have data in a structured
format but now as the volume of the data is increasing,
Why machine
Learning becomes
so the number of structured data becomes very less,
more popular these so to handle the massive amount of data we need data
days?
science techniques
• Those data can be used to get the proper business
insights and the hidden trends from them.
• These insights helps the organization to predict the
Future
• Helps to reduce the production cost
• Build model based on the data to give the ability to the
machine to predicts on its own
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
What is Machine Learning?
Machine learning is a sub-set of artificial intelligence (AI) that allows the system to
automatically learn and improve from experience without being explicitly programmed
Training Data Model Building Testing Data
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Traditional Vs Machine Learning
Traditional Programming Machine Learning
Data Data
Model
Output
Program Output
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
What is Machine Learning?
Process to train a machine Learning model
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Life cycle To build a model with Machine Learning
Understand the business
Data Acquisition Data Cleaning
problem
Exploratory data Analysis
Predict your model
Deploy the model Machine Learning Algorithm
accuracy
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
Types Of Machine Learning
Supervised Learning
Unsupervised
Learning
Reinforcement
Learning
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
What is Supervised Learning?
Supervised learning works as a supervisor or teacher. Basically, In supervised learning, we teach or train the
machine with labeled data (that means data is already tagged with some predefined class). Then we test our
model with some unknown new set of data and predict the level for them
Learning from the labelled data and applying the knowledge to predict the
label of the new data(test data), is known as Supervised Learning
Types of Supervised Learning:
Regression
Classification
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
What Is Classification?
“Classification is the process of grouping similar data points”
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Classification Vs Regression
Classification
What will be the weather for tomorrow?
rainy or sunny?
Predicts categorical outcome
Regression
Predicts continuous outcome
What is the temperature forecasting for
tomorrow?
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Types Of Classification Algorithm
Logistic Regression
Decision Tree
Random Forest
K- Nearest Neighbor
Naïve Bayes
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Naïve Bayes Classification
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Naïve Bayes
“Naive Bayes classifiers are belonging from simple "probabilistic classifiers" based on Bayes'
theorem ”
• P(B|A) = Probability of B being true
P(B|A)P(A)
P(A|B) = given that A is true
P(B)
• P(A) = Probability of A being true
• P(B) = Probability of B being true
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
What Is A Naïve Bayes Algorithm?
“ Naive Bayes classifier assumes that the presence of a particular feature in a class is
unrelated or independent to the presence of any other features ”
For Example:
Colour: RED
Shape: ROUND
Diameter: 3 inches
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Why Is Naïve Bayes Algorithm Called So?
Q1: Why was the Algorithm named as Naïve?
• Assumptions are taken by the algorithms are very naïve
• Those assumptions can may or may not be correct.
Q2: Why the Algorithm has the word Bayes in it?
Fundamentals are based on the Bayes Theorem.
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Bayes Theorem: Example
What is the probability that your flight will be on time today given that
weather is sunny and humidity is low?
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Bayes Theorem: Example
For example:
Finding the probability of having a disease given you were tested
positive
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Conditional Probability
Define 2 events…
Those two events are related to each other
Event A is the probability of the event we’re trying to calculate
Event B is the condition that we know or the event that has happened
Conditional Probability : P(A|B) ,
The probability of the occurrence of event A given that B has already happened
P(A|B)
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Understanding Conditional Probability: Example
• Q: Probability of getting 1 by rolling a dice.
• Given: After rolled the dice it is an odd number
Solution:
Event A = Getting 1 by rolling the dice
Event B = Its is an odd number
P(A) = 1/6
After given the condition
Possibility of getting an odd dice = 3 (1,3,5)
P(A|B) = 1/3
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Understanding Bayes Theorem
• The Bayes theorem describes the probability of an event based on the prior knowledge of the
conditions that might be related to the event.
• In simple words,
• Bayes' theorem shows the relation between a conditional probability and its reverse form.
• If conditional probability = P(A|B)
• You can use Bayes rule to find the reverse probabilities i.e. P(B|A)
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Explanation of Bayes Theorem:
This is likelihood
of seeing that
This is prior i.e what
evidence if your
you believed before
hypothesis is
you saw the
correct
evidence
This is likelihood of
This is that evidence
posterior under any
circumstances
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Bayes Theorem: Example
Find out the probability of having liver disease if the patient is an
alcoholic
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Summarizing The Dataset
On time
Flight will on time or not?
D1
D2
D3
D4 Total Sample: 14
D5
No. of Attributes: 4
D6
D7 No. of Class: 1
D8
D9
Total Yes: 9 P(Yes): 9/14
D10
D11 Total No: 5 P(No): 5/14
D12
D13
D14
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Diabetes Prediction using Naïve Bayes
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
Logistic Regression
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
What is Logistic Regression?
Logistic regression is also a part of supervised learning classification algorithm. It is used to predict the
probability of a target variable and the nature of target or dependent variable is discrete, so for the
output there will be only two class will be present
The dependent variable is binary in nature so that can be either 1 (stands for
success/yes) or 0 (stands for failure/no).
Logistic regression is also known as sigmoid function
Sigmoid function = 1 / (1 + e^-value)
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
When Logistic Regression comes into picture?
Y
Size of Dog (in inch) Linear Regression
size = 2 . 1 x weight + 4
But, if weight = 0 kg
Predicted Size = 4 inches
HOW is it even possible?
X
Weight of Dog
(in kg)
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
When Logistic Regression comes into picture?
These dogs are obese..
Probability
Dog is Obese
These are not..
Dog is NOT Obese
Weight
Logistic Regression
Similar to linear regression except, it predicts whether something is TRUE or FALSE, instead predicting something continuous like
size of dog
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Understanding of Logistic Regression
Probability
High probability of the new dog being an
obese
Weight
Logistic Regression
The curve tells the probability of the dog being an obese, based on its weight.
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Understanding of Logistic Regression
Probability
50% chance that the dog is
obese
Weight
Logistic Regression
The curve tells the probability of the dog being an obese, based on its weight.
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Understanding of Logistic Regression
Probability
Very low probability of a light dog being an
obese
Weight
Logistic Regression
It is generally used for classification.
For example, If the probability of the dog being obese is > 50% then it is classified as ‘obese’
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
or else ‘not obese’
Mathematical Explanation of Logistic Regression
Logistic Regression is a special case of Linear Regression
Linear Regression Sigmoid Function
1
Y= 𝑚𝑚1𝑥𝑥1 + 𝑚𝑚2 𝑥𝑥2 + ⋯ + 𝑚𝑚1𝑥𝑥n + 𝑐𝑐 𝑝𝑝 =
1 + 𝑒𝑒 −𝑦𝑦
• Y dependent variable and
Logistic Regression
• 𝑥𝑥1, 𝑥𝑥2 ... 𝑥𝑥𝑛𝑛 independent/explanatory
1
variables 𝑝𝑝 =
1 + 𝑒𝑒 −(𝑚𝑚1 𝑥𝑥1 + 𝑚𝑚2𝑥𝑥2 +⋯+ 𝑚𝑚1 𝑥𝑥n +𝑐𝑐)
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Real time Use cases that can come under Logistic Regression Problem
Will the property have a good neighborhood?
The mail is spam mail or not?
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Key points of Logistic Regression
LOGISTIC REGRESSION
Categorical Variables
Solves Classification Issue
S-Curve
Eg: Is the mail spam or not
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Credit Card Fraud Analysis Using Logistic Regression
Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited.
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
Thank You
Proprietary content. ©Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited
