Machine Learning

with Python
Machine Learning
Boot camp
Presented by
R Sreenivas
M Tech
RGUKT RK Valley
 Agenda

01 Introduction to ML

02 Types of Learning

03 Steps to build ML

04 Python Libraries for ML

Artificial Intelligence (AI)
AI is a branch of computer science by which we
can create intelligent machines which can behave
likes a human, think like humans, and able to
make decisions.
AI exists when a machine can have human skills
such as learning, reasoning as solving problems.

AI
 Artificial Intelligence
Natural
Computer Language
Vision & Processing & Machine
Deep Speech Learning &
Robotics Sensors Data Science
Learning Technology

Human Human Human Human Human

Can Can Can Can can
Move See & Sense Listen & Speak Learn &
Recognize Analysis
 Machine Learning

What How Why Where

? ? ? ?

What is ML ? How to Learn ? Why we Learn ? Applications

Definition Programming Job Weather Prediction

Types Skills Social Media Analysis
Autonomous Vehicle
Steps Search Engine
Diseases detection
Diseases classifications
Machine Learning
Machine Learning is a concept which allows the machine
to learn from the experience using algorithms, and that too
without being explicitly programmed. ~ Arthur Samuel (1959)

A computer program is said to learn “ if its performance

at a task T , as measured by a performance P, improves with
experience E." ~Tom Mitchell(1997) ML
Types of Learning

Unsupervised Learning Reinforcement Learning

 Supervised Learning
Supervised learning is the types of machine learning in which machines are
trained using well "labelled" training data, and on basis of that data, machines
predict the output

Supervised
Learning
 Supervised Learning
Classification Regression
A classification problem is when A regression problem is when the
the output variable is a categorical output variable is a real or
Algorithms : Logistic regression, Decision tree,
continuous value.
KNN .. etc
Algorithms : Linear regression, Polynomial
regression .. etc
 Unsupervised Learning
Unsupervised learning is a type of machine learning in which models are
trained using unlabeled dataset . The goal of unsupervised learning is to find
the underlying structure of dataset, group that data according to similarities

Unsupervised
Learning
 Unsupervised Learning
Clustering Association
Clustering is the method of grouping Association is the learning method
the objects into clusters such that object
which is used for finding the relationship
with most similarities remains into a same
between the variables .
group and has less or no similarities with the
object of another group .
Algorithms : K-Means clustering , SVD ..etc
 Reinforcement Learning
Reinforcement learning is a feed back based machine learning technique
in which an agent learns to behave in an environment by performing actions ,
on seeing the result of actions.

Reinforcement
Learning
1 Data Collection

Data Preparation 2
3 Choose a model

Training 4

5 Evaluation

Parameter tuning 6

7 Prediction
 Data Collection
Data collection means pooling data by scraping, capturing, and
loading it from multiple sources, including offline and online
sources.

Sepal Length Sepal Width Petal Length Petal Width

Species
Cm Cm Cm Cm
5.1 3.5 1.4 0.2 Iris-setosa
4.9 3.0 1.4 0.2 Iris-setosa
7.0 3.2 4.7 1.4 Iris-versicolor
6.4 3.2 4.5 1.5 Iris-versicolor
6.5 3.0 5.2 2.0 Iris-virginica Data
5.9 3.0 5.1 1.8 Iris-virginica
50 2.0 - 5.0 Iris-virginica Collection
Data Preparation
Raw data alone is not very useful. The data needs to be prepared,
normalized, de-duplicated and errors and bias need to be removed.
Visualisation of the data can be used to look for patterns and outliers
to see if the right data has been collected or if data is missing.

Sepal Length Sepal Width Petal Length Petal Width

Species
Cm Cm Cm Cm
5.1 3.5 1.4 0.2 Iris-setosa
4.9 3.0 1.4 0.2 Iris-setosa
7.0 3.2 4.7 1.4 Iris-versicolor
6.4 3.2 4.5 1.5 Iris-versicolor Data
6.5 3.0 5.2 2.0 Iris-virginica
5.9 3.0 5.1 1.8 Iris-virginica
Preparation
Choose a model
Model selection is a technique for selecting the best model after
the individual models are evaluated based on the required
criteria. There are many models that can be used for many different
purposes.

Models
 Logistic Regression
 Decision Tree
 KNN
 Support Vector Machine
 Linear Regression
Choose a
 K – Means ...etc Model
Training
The process of training an ML model involves providing training data
to ML algorithm (that is, the learning algorithm) , then model learn
from it.
Sepal Length Cm Sepal Width Cm Petal Length Cm Petal Width Cm

Sepal Sepal Petal Petal 4.9 3.0 1.4 0.2

Length Width Length Width Species
Cm Cm Cm Cm 6.4 3.2 4.5 1.5
0
5.1 3.5 1.4 0.2
1 Species
7.0 3.2 4.7 1.4
2 0
6.5 3.0 5.2 2.0
Logistic Predictive 1
5.9 3.0 5.1 1.8 2
Regression model

Training
Evaluation
Model Evaluation is the process through which we quantify the
quality of a system’s predictions. It means that check the model
accuracy.

For Classification -> Precision, Recall, Accuracy, F1-score …etc

For Regression -> MSE, MAE, R-Squared …etc

Evaluation
Parameter tuning
After evaluating your model, we should test the originally set
parameters to improve the AI. Increasing the number of training
cycles can lead to more accurate results.

Parameter
tuning
Prediction
Once we have gone through the process of collecting data,
preparing the data, selecting the model, training and evaluating the
model and tuning the parameters , our model is ready to predict real
time data.

Prediction
NumPY Pandas Scikit-learn Matplotlib

ML ML ML ML

Scipy Seaborn Theano

ML ML ML

TensorFlow Keras
ML ML
 NumPY
ML
 NumPy stands for Numerical Python
 NumPy is a popular Python library for multi-dimensional
array and matrix processing because it can be used to ML
perform a great variety of mathematical operations.
 Its capability to handle linear algebra, Fourier transform, and
ML
more, makes NumPy ideal for machine learning and artificial
intelligence (AI)
 Pandas
ML
 Pandas is the Python data manipulation and analysis library.
 Pandas is responsible for preparing data sets and points for
machine training. ML
 Pandas uses two types of data structures, one-dimensional
(series) and two-dimensional (DataFrame), which, together,
ML
allow Pandas to be used in a variety of sectors, from science
and statistics to finance and engineering.
 Scikit-learn
ML
 Scikit-learn is a very popular machine learning library that is
built on NumPy and SciPy.
 It supports most of the classic supervised and unsupervised ML
learning algorithms, and it can also be used for data mining,
modeling, and analysis.
ML
 Scikit-learn offers a user-friendly library for those new to
machine learning.
 Matplotlib
ML
 Matplotlib is a Python library focused on data visualization
and primarily used for creating beautiful graphs, plots,
histograms, and bar charts. ML
 t is compatible for plotting data from SciPy, NumPy, and
Pandas.
ML
 Matplotlib is intuitive and easy to use, making it a great
choice for beginners.
 Scipy
ML
 SciPy is a free and open-source library that’s used to perform
scientific and technical computing on large sets of data.
 SciPy comes with embedded modules for array optimization ML
and linear algebra.
 SciPy is ideal for image manipulation and provides basic
ML
processing features of non-scientific high-level mathematical
functions.
 Seaborn
ML
 Seaborn is an open-source data visualization and plotting
Python library. It’s based on the plotting library Matplotlib
and includes the extensive data structures of Pandas ML
 Seaborn produces the most visually appealing and attractive
graphs and plots, making it perfect for use in publications and
ML
marketing.
 Theano
ML
 Theano is a numerical computation Python library made
specifically for machine learning.
 It is able to optimize and evaluate mathematical models and ML
matrix calculations that use multi-dimensional arrays to
create ML models.
ML
 TensorFlow
ML
 TensorFlow is a free and open-source Python library that
specializes in differentiable programming.
 TensorFlow can be used to implement reinforcement- ML
learning in ML and DL models and allows you to directly
visualize your machine learning models with its built-in tools.
ML
 Keras
ML
 Keras is an open-source Python library designed for
developing and evaluating neural networks within deep
learning and machine learning models. ML
 Keras is flexible, portable, and user-friendly, and easily
integrated with multiple functions.
ML
 Importing necessary packages
Data preparation and preprocessing
Segregation of Data (Independent and Dependents)
Splitting the dataset into train data and test data
Choosing the model
Training the model
Testing model
Evaluation of the model
Prediction
 Importing necessary packages

import numpy as np ML

import pandas as pd
import matplotlib.pyplot as plt ML
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression ML

from sklearn import metrics

 Data preparation and preprocessing

years_of_exp = [2,4,6,8,10,12,14,16,18,20] ML

salary=[600000,800000,1000000,1200000,1400000,1600000,
1800000,2000000,2200000,2400000] ML
dataset = pd.DataFrame({"Year of Experice":years_of_exp,
"Salary":salary})
print(dataset) ML

sns.pairplot(dataset)
 Data preparation and preprocessing(Contd.)

Year of Experice Salary ML

2 600000
4 800000
6 1000000
ML
8 1200000
10 1400000
12 1600000
14 1800000
16 2000000
ML
18 2200000
20 2400000
 Segregation of Data (Independent and Dependents)
Year of ML
X = dataset.drop('Salary',axis=1) Salary
Experience
2 600000
Y = dataset['Salary']
4 800000
6 1000000
ML
8 1200000
10 1400000
12 1600000
14 1800000
2000000
ML
16
18 2200000
20 2400000
 Splitting the dataset into train data and test data

from sklearn.model_selection import train_test_split ML

X_train,X_test,Y_train,Y_test = train_test_split(X,Y,test_size=0.2)
Year of Salary
Experience ML
2 600000
4 800000
6 1000000
8 1200000
10 Year of 1400000 ML
Salary
12 Experience 1600000
18 1800000 2200000
14
20 2000000 2400000
16

X_train X_test Y_train Y_test

 Choosing the model

from sklearn.linear_model import LinearRegression ML

model = LinearRegression()

ML

Linear
Regression
ML
 Training the model
ML
model.fit(X_train,Y_train)

Year of
Experience
Salary ML
2 600000
4 800000
6 1000000 Linear
8 1200000 Regression
10 1400000 ML
12 1600000
14 1800000
16 2000000

X_train Y_train
 Testing the model

predictions = model.predict(X_test) ML

print(predictions)

ML
Year of Salary
Experience Linear 2200000
18
20
Regression 2400000
ML
X_test Predictions
 Evaluation of the model

from sklearn import metrics ML

print(metrics.mean_squared_error(Y_test,predictions))
print(metrics.mean_absolute_error(Y_test,predictions))
ML
print(np.sqrt(metrics.mean_squared_error(Y_test,predictions)))
Salary Salary
2200000 2200000 ML
2400000 2400000

Y_test Predictions
 Predictions
experience = np.array([[int(input("Enter your years of experince :"))]]) ML
salary = model.predict(experience.reshape((-1,1)))
print(salary)
ML
Year of Salary
experience Linear 10400000
100 Regression
ML
new data
Predictive
model
 Model Development

 Model development is creating a mathematical ML

representation by generalizing and learning from

training data that is used for predictions. ML
 sklearn – This is the module contains multiple
sub-modules having pre-implemented functions to
perform tasks from data preprocessing to model ML

development and evaluation.

 Linear Regression

 Linear Regression is a supervised learning algorithm. ML

 It is used to solve regression problems.

 It finds the relationship between the dependent(Y) and ML
independent variable(X).
 It finds the best fit linear line between the independent
and dependent features. ML
Linear Regression
ML

ML

Regression line ML
Data point
 Types of Linear Regression
ML
Simple Linear Regression Multiple Linear Regression

Attended Attended Books ML

Marks (Y) Marks (Y)
class (X) class (X1) Referred (X2)
3 8 3 4 8
4 9 4 5 9
3 7 3 3 7 ML
5 10 5 5 10
2 6 2 3 6
 Simple Linear Regression

 Simple linear regression is a regression algorithm. ML

 It is used when a dataset contains only one independent

variable(X) and a dependent variable(Y).
ML
 Mathematical equation  y=a+bx
where ,
y is dependent variable
ML
x is independent variable
b is coefficient of x
a is ‘y’ intercept
 Simple Linear Regression
a) Find regression line equation for below data ML
b) Predict the marks of the student if he/she attended 10
classes.
Attended Marks (Y) ML
y=a+bx class (X)
3 8
4 9
∑Y = N a + b ∑X 3 7
5 10
ML
∑XY = a ∑X + b ∑X^2 2 6
 Simple Linear Regression
a) Find regression line equation for below data ML
b) Predict the marks of the student if he/she attended 10
classes.
Attended class Marks (Y) XY X^2 Attended Marks (Y) ML
(X) class (X)
3 8 3 8
4 9 4 9
3 7 3 7
5 10 5 10
ML
2 6 2 6
∑X= ∑Y= ∑ XY = ∑ X^2 =
 Simple Linear Regression
a) Find regression line equation for below data ML
b) Predict the marks of the student if he/she attended 10
classes.
Attended Marks (Y)
∑Y = N a + b ∑X ML
class (X)
∑XY = a ∑X + b ∑X^2 3 8
4 9
a = 3.4236 3 7
b = 1.346 5 10 ML
y=a+bx 2 6
y = 3.4236 + 1.346 x
 Simple Linear Regression
Evaluation of model ML
1) Mean Squared Error ( MSE ) = 1/N * ∑ ( y – ŷ )^2
2) Mean Absolute Error ( MAE ) = 1/N * ∑ | y – ŷ |
3) R - Squared Error ( RSE ) = 1 – ( ∑ ( y – ŷ )^2 / ∑ ( y – y )^2 )
Attended Marks ML
class (X) (Y)
ŷ Y–ŷ ( y – ŷ )^2 |y–ŷ| ( y – y )^2
3 8
4 9
3 7 ML
5 10
2 6
∑Y ∑( y – ŷ )^2 ∑| y – ŷ | ∑ ( y – y )^2
 Multiple Linear Regression

 Multiple linear regression is a regression algorithm. ML

 It is used when dataset contains more than one

independent variables and a dependent variable
ML
 Mathematical equation  y = a + b1 x1 + b2x2 + …..
where ,
y is dependent variable
ML
x1,x2,… are independent variables
b1,b2,…. is coefficient of x1,x2,..
a is ‘y’ intercept
 Multiple Linear Regression
a) Find the regression line equation for below data ML
b) Predict the marks of the student if x1 = 10 , x2 = 9
Attende Books Marks
y = a + b1 x1 + b2x2 d class Referred (Y)
(X1) (X2)
ML
3 4 8
∑Y = N a + b1 ∑X1 + b2 ∑X2 4 5 9
3 3 7
∑YX1 = a ∑X1 + b1 ∑X1^2 + b2 ∑X1X2 5 5 10 ML
∑YX2 = a ∑X2 + b1 ∑X1X2 + b2 ∑X2^2 2 3 6
 Multiple Linear Regression
a) Find the regression line equation for below data ML
b) Predict the marks of the student if x1 = 10 , x2 = 9

Attende Books Marks YX1 YX2 X1X2 X^2 X^2

d class referred (Y)
ML
(X1) (X2)
3 4 8
4 5 9
3 3 7
5 5 10 ML
2 3 6
∑X1 = ∑X2 = ∑Y = ∑YX1 = ∑YX2 = ∑X1X2 = ∑X1^2 = ∑X2^2 =
 Multiple Linear Regression
a) Find the regression line equation for below data ML
b) Predict the marks of the student if x1 = 10 , x2 = 9

∑Y = N a + b1 ∑X1 + b2 ∑X2
Attended Marks (Y)
ML
∑YX1 = a ∑X1 + b1 ∑X1^2 + b2 ∑X1X2 class (X)
3 8
∑YX2 = a ∑X2 + b1 ∑X1X2 + b2 ∑X2^2 4 9
a = 2.52 3 7
5 10 ML
b1 = 0.833 , b2 = 0.667
y = a + b1x1 + b2x2 2 6
y = 2.52 + 0.833 x1 + 0.667 x2
 Multiple Linear Regression
Evaluation of model ML
1) Mean Squared Error ( MSE ) = 1/N * ∑ ( y – ŷ )^2
2) Mean Absolute Error ( MAE ) = 1/N * ∑ | y – ŷ |
3) R - Squared Error ( RSE ) = 1 – ( ∑ ( y – ŷ )^2 / ∑ ( y – y )^2 )
Attended Marks ML
class (X) (Y)
ŷ Y–ŷ ( y – ŷ )^2 |y–ŷ| ( y – y )^2
3 8
4 9
3 7 ML
5 10
2 6
∑Y ∑( y – ŷ )^2 ∑| y – ŷ | ∑ ( y – y )^2