Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn

Machine Learning Example
How does
Snapchat
filter work?

Matrix of numbers

Matrix of numbers
Looks for color patterns
on face

Matrix of numbers
Looks for color patterns
on face
Points align
themselves and
look for areas of
contrast

What’s in it for you?
Types of Machine Learning Algorithms
Real World Applications of Machine Learning
Processes involved in Machine Learning
What is Machine Learning?
Logistic Regression
Linear Regression
Decision Tree and Random forest
K Nearest Neighbors
Popular Algorithms with Hands On Demo

Real World Applications of
Machine Learning

Real World Applications of Machine Learning
Face Recognition Healthcare Industry Weather Forecasting
Produce a Web Series Prepare a new Drink
Siri and Cortana

Machine Learning is the science of making computers learn and act like humans by feeding
data and information without being explicitly programmed.

Past Data

Past Data
Analyses

Past Data
Analyses
Trains

Past Data
Analyses
Predicts
Trains
Output

Data
Gathering
7Machine Learning
Processes

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing
Choose
Model

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing
Choose
Model
Train Model

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing
Choose
Model
Train Model
Test Model

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing
Choose
Model
Train Model
Test Model
Tune Model

Data
Gathering
7Machine Learning
Processes
Data
Pre-Processing
Choose
Model
Train Model
Test Model
Tune Model
Prediction

Types of Machine Learning Algorithms
VSupervised
Learning
Un-Supervised
Learning
Reinforcement
Learning
Machine
Learning
Association
• Market Basket Analysis
• Text Mining
• Face Recognition
Classification
• Fraud Detection
• Email Spam Detection
• Image Classification
Regression
• Weather Forecasting
• Risk Assessment
• Score Prediction
Clustering
• Medical Research
• City Planning
• Targeted Marketing
Reinforcement Learning
• Gaming
• Robot Navigation
• Stock Trading
• Assembly Line Processes

Popular Algorithms in Machine Learning
Algorithm 1
Linear
Regression

Algorithm 1
Algorithm 2
Linear
Regression
Logistic
Regression

Algorithm 1
Decision
Tree
Algorithm 2
Algorithm 3
Linear
Regression
Logistic
Regression

Algorithm 1
Decision
Tree
Random
Forest
Algorithm 2
Algorithm 3
Algorithm 4
Linear
Regression
Logistic
Regression

Algorithm 1
Decision
Tree
Random
Forest
K Nearest
Neighbors
Algorithm 2
Algorithm 3
Algorithm 4
Algorithm 5
Linear
Regression
Logistic
Regression

History of Linear Regression
Linear Regression: Brief history on how Regression came into picture.
Francis Galton Heights of Father and Son
Studied
Best fit
line
Regression Line
Analyzed
Height Data
Mean Height of all People
Regressed

Linear Regression
Linear Regression is a linear modelling approach to find relationship between one or more independent variables (predictors)
denoted as X and dependent variable (target) denoted as Y.
Temperature
Sales
Temperature
Sales
Regression line to predict the sales
Plotting the sales of ice cream based
on temperature
Predict sales of Ice Cream based on temperature:

Linear Regression
Finding the best fit line: The best fit line can be found out by minimizing the distance between all the data points and the distance
to the regression line. Ways to minimize this distance are sum of squared errors, sum of absolute errors etc.
Distance between the data points
and the line is maximum
Not the best fit line

Linear Regression
Not the best fit line
and the line can still be reduced

Linear Regression
and the line is the least
Best fit regression line

Implementation of Linear Regression
Linear Regression: Predict Employee Salary based on Years of Experience.
HR team trying to figure out how much
to pay to a new joinee?
Predict Salary
How much do we pay her?
New Joinee
Years of experience
Salary
Employee data of Salary
based on Years of experience

1. Load the libraries:
2. Import the Dataset:

3. Visualize the data:

4. Split the data into training and testing set:
5. Fit Simple Linear Regression to the training dataset:

6. Predict the test set results:

7. Visualize the train set results:

8. Visualize the test set results:

9. Calculating the residuals:
Since the RMSE value is too less, the model is a good model

Logistic Regression
Logistic Regression is a Classification algorithm used to predict discrete/ categorical values
Who will default on their credit card payment?

Logistic Regression
Credit card users

Logistic Regression
Credit card users Make credit card
transaction

Logistic Regression
Credit card users
Plot of monthly credit card balance
and annual income
Make credit card
transaction

Logistic Regression
Plotting the Logistic Regression Curve: The Logistic Regression curve is known as the Sigmoid curve (S curve)
Predicted Y lies in the range 0 and 1
1
1 + e-zP =
Predicted Y can exceed the range 0 and 1
Y = m1x+c0

Logistic Regression
Plotting the Logistic Regression Curve: The Logistic Regression curve is known as the Sigmoid curve (S curve)
Y = m1x+c0
0.5
Threshold
value
0.5
Cutoff point at 0.5, anything below it
results in 0 and above is 1
Red data point will default as it is above
the threshold value of 0.5 and green data
point won’t as it is below the threshold
value

Implementation of Logistic Regression
Logistic Regression: Predict if a person will buy an SUV based on their Age and Estimated Salary
Age Estimated
Salary

2. Import the dataset and extract the independent and dependent variables:

3. Visualize the dataset:

4. Split the dataset into Training and Testing set:
5. Feature Scaling:
6. Fit Logistic Regression to Training dataset:

7. Predicting the Test set results:

8. Visualize the Train set results:

9. Visualize the Test set results:

11. Evaluating the model:
Understanding the Confusion Matrix:
N = 134 Predicted: No Predicted: Yes
Actual: No TN=79 FP=6
Actual: Yes FN=11 TP=38
Accuracy:
(TN+TP)/N
(79+38)/134 = 0.87
Misclassification Rate:
(FP+FN)/N
(6+11)/134 = 0.13

Decision Tree and
Random Forest

Decision Tree
Decision Tree is a tree where each node represents a feature (attribute), each link (branch) represents a decision and each leaf
represents an outcome (categorical or continuous value).
Should the person accept a new job offer?
Yes
Salary >
$60,000
Root Node
Accept a job offer?
Reject
Offer
No
Leaf node
Accept
offer
Yes
Reject
Offer
No
Reject
Offer
YesNo
Performance
Incentives
Decision
node
Commute > 1
hour

Implementation of Decision Tree
Decision Tree and Random Forest: Does Kyphosis exist after a surgery?
Bunch of kids Kyphosis surgery Predict kyphosis
present or absent?

Classification Tree for Kyphosis:
Start>=8.5
Absent 64/17
Present
8/11
No
Start>=14.5
Absent 56/6
Yes
Absent
12/0
Yes
Absent
29/0
Yes
Absent
12/2
Yes
Age<55
Absent 27/6
No
Age>=111
Absent 15/6
No
Present
3/4
No

2. Import the dataset and extract the independent and dependent variables:

3. Visualize the data:

4. Split the data into Train and Test set:
5. Train a Decision Tree:

6. Predict the Model:

7. Evaluate the Model:
Understanding the confusion matrix: Accuracy:
(TN+TP)/N
(16+3)/25 = 0.76
(FP+FN)/N
(4+2)/25 = 0.24

Improving performance using Random Forest
Compare the model using Random Forest Tree:
Accuracy:
(TN+TP)/N
(18+2)/25 = 0.80
(FP+FN)/N
(2+3)/25 = 0.20

Algorithm 5
K Nearest Neighbors

K Nearest Neighbors (KNN)
K Nearest Neighbors: KNN is a Classification algorithm generally used to predict categorical values.
Height(ft)
Weight (lbs)
Class Dog
1.5
3.5
2.0
3.0
2.5
1.0
0.5
0 65 75 9585 105
Class Cat
Weight (lbs)
Class Dog
1.5
3.5
2.0
3.0
2.5
1.0
0.5
0 65 75 9585 105
Height(ft)
New data
point
Class Cat
To find if a new data point is a or a ?

Choosing a K will define what class a new data point is assigned to:
K Nearest Neighbors
Height(ft)
Class Dog
Weight (lbs)
0.5
2.5
2.0
1.5
3.0
1.0
3.5
10565 75 85 95
Class Cat
K=3
If K=3, the new data point
belongs to class Cat
K=7
If K=7, the new data point
belongs to class Dog

Implementation of KNN
Predict if a person will buy an SUV based on Age and Estimated Salary
Age Estimated
Salary

1. Load the Libraries:
2. Load the Dataset:

3. Visualize the dataset:

4. Split the data into Train and Test set:
5. Feature Scaling:

6. Fit KNN to Train set:
7. Predict the Test set results:

8. Visualize the Train set results:

9. Visualize the Test set results:

10. Evaluate the model by creating a confusion matrix:
Understanding the confusion matrix:
Accuracy:
(TN+TP)/N
(64+29)/100 = 0.93
(FP+FN)/N
(4+3)/100 = 0.07

Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn

Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn

Similar to Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn (20)

More from Simplilearn

More from Simplilearn (20)

Recently uploaded

Recently uploaded (20)

Machine Learning Algorithms | Machine Learning Tutorial | Data Science Algorithms | Simplilearn

Editor's Notes