Overview of AI and ML Terminology awareness Applications in real world Use cases within Nokia Types of Learning Regression Classification Clustering Linear Regression Single Variable with python

1. Machine
Learning?
What
is
Soumya Mukherjee
Md Shamimuddin
https://www.linkedin.com/in/soumyarmukherjee/
https://www.linkedin.com/in/mdshamimuddin/

2. Agenda
 Overview of AI and ML
 Terminology awareness
 Applications in real world
 Use cases within Nokia
 Types of Learning
 Regression
 Classification
 Clustering
 Linear Regression Single Variable with python

3. • Arthur Samuel (1959)
Machine Learning: Field of study that gives computers the
ability to learn without being explicitly programmed.
• Tom Mitchell (1998)
A computer program is said to learn from experience E with
respect to some task T and some performance measure P, if its
performance on T, as measured by P, improves with experience E.
Machine Learning Definition

4. Artificial Intelligence Vs Machine Learning Vs Deep Learning

5. Terminology Awareness

6. Implies huge data
volumes that cannot be
processed effectively with
traditional applications.
Big Data processing
begins with raw data that
is not aggregated and it is
often impossible to store
such data in the memory
of a single computer
Is about using Statistics
as well as other
programming methods to
find patterns hidden in
the data so that you can
explain some
phenomenon. Machine
Learning uses Data
Mining techniques and
other learning algorithms
to build models of what is
happening behind
some data.
Big Data Data Mining
Is an artificial
intelligence technique
that is broadly used
in Data Mining. ML uses
a training dataset to build
a model that can predict
values of target variables.
Data Mining uses the
predictive force of
Machine Learning by
applying various ML
algorithms on Big data.
Machine Learning

7. WHAT IS ARTIFICIAL INTELLIGENCE
• Artificial intelligence (AI) is an area of computer science that emphasizes the creation of intelligent
machines that work and react like humans. Some of the activities computers with artificial intelligence
are designed for include:
Knowledge
Gain
Reasoning
Problem
Solving
Learning

8. Artificial Intelligence
Machine Learning
Supervised Unsupervised Reinforcement

9. Types of Learning
Supervised
Learning
Unsupervised
Learning
Reinforcement
Learning
Target/outcome
variable to be
predicted from set of
predictors is known
at training phase.
E.g. Regression,
Decision Tree,
Random Forest, KNN
Target/outcome
variable to be
predicted from set of
predictors is
unknown at training
phase.
E.g. Clustering (K-
means, Apriori)
Machine is trained to
take specific decision
Exposed to an
environment where it
trains itself
continually using trial
and error.
E.g. Markov Decision
process

10. Applications in real world
• Google search engine
• Self driving cars
• Facebook auto tagging
• Netflix movie recommendation
• Amazon product recommendation
• Healthcare diagnosis
• Speech recognition
• StackOverflow QA tagging
• Chatbot

11. Data as input
(Text files,
spreadsheet,
SQL database)
Feature Engineering
(Removing unwanted data,
Handle missing values,
Normalization or
Standardization)
Algorithm
Output/
Model
Pipeline solving ML Problem

12. Pipeline in solving ML Problem

13. Data Exploration/Feature Engineering
1. Variable Identification
• Predictor(s) n Target
• Type n Category of variable
2. Univariate Analysis
• Central tendency
• Measure of Dispersion
• Visualization Method
• Frequency table(categorical)
3. Bivariate Analysis
• Relation between 2 variables
• Correlation
• Chi-square test
• Z-test
4. Missing Value
Treatment
• Deletion
• Imputation
• Prediction Model
• KNN Imputation
5. Outlier Handling
Detection
• Very Important to handle outlier
• Visualization technique like box-
plot, scatter plot, Histogram
• Any value beyond -1.5IQR to
1.5IQR is an outlier
Treatment
• Remove
• Scale or Normalize
• Transform
• Impute

14. SUPERVISED LEARNING
• Supervised learning is used whenever we want to predict a certain outcome from
a given input, and we have examples of input/output pairs.
• We build a machine learning model from these input/output pairs, which
comprise our training set.
• Our goal is to make accurate predictions for new, never-before-seen data.
• Supervised learning often requires human effort to build the training set, but
afterward automates and often speeds up an otherwise laborious or infeasible
task.

15. TYPES OF SUPERVISED MODEL
• Regression :
• regression is the process of predicting a continuous value
• Classification
• predict a class label, which is a choice from a predefined list of possibilities.

16. CLASSIFICATION
• Binary Classification : Distinguishing between exactly two classes
• Multiclass classification : Classification between more than two classes.

17. Types of regression
1. Simple Linear Regression
Single predictor + single target
y = m*x + c
2. Multiple Linear Regression
Multiple predictors + single target
y = m1*x1 + m2*x2 + c
3. Polynomial Regression
One or many predictors + single target
Y = mn * x^n + … + m2*x^2 + m1*x1 + c
4. Stepwise Regression
Useful in case of multiple predictors
Add or Remove predictors as needed
Forward selection
Backward elimination
5. Lasso Regression
6. Ridge Regression
7. ElasticNet Regression

18. Simple Linear Regression
• Single predictor and single target
• Y = b0 + b1*X
• Minimum sum squared error
• Standard packages are already available
• Formula
• Programming example

19. Classification
 Type of supervised learning
 Output or target is a categorical outcome
Example
 Mail spam or no spam
 Weather rainy, sunny, humid
 Stock price up or down
Predictor(s) Algorithm
Categorical
Target

20. Types of Classification
1. K-nearest Neighbor Classifier
2. Logistic Regression
3. Naïve Bayes 6. Support Vector Machine
Classifier
5. Random Forest Classifier
4. Decision Tree Classifier

21. Clustering (Unsupervised learning)
Cluster 1
Cluster 2
Cluster 3

22. Unsupervised learning
• Unsupervised learning is the training of machine using
information that is neither classified nor labelled
For instance, Given an image having both dogs and cats which have not seen ever.
Machine tries to find pattern
based on shape of head,
ears, body structure etc.

23. Reinforcement Learning
• Reinforcement learning (RL) is an area of machine learning concerned with
how software agents ought to take actions in an environment so as to maximize some
notion of cumulative reward. (source : Wikipedia)
Eg : you go near fire , its warm : positive reinforcement
you touch fire, it burns your hand : negative reinforcement  learn not to touch
fire
• Algorithms for RL include – MonteCarlo methods, Markov Decision Processes, Q-
learning etc

24. ML in Python:
• Numpy
• Pandas
• Scikit-learn
• Matplotlib
• Seaborn
Non-
Programming:
• Weka
• Orange
• RapidMiner
• Qlik Sense
• xls
Deep Learning:
• Tensorflow
• Keras
• PyTorch
• Theano
Tools And Packages

25. LINEAR REGRESSION
SINGLE VARIABLE

26. LINEAR REGRESSION
• Linear regression, or ordinary least squares (OLS), is the simplest and most classic
linear method for regression. Linear regression finds the parameters m and b that
minimize the mean squared error between predictions and the true regression
targets, y, on the training set.

27. HOME PRICES
area price
2600 550000
3000 565000
3200 610000
3600 680000
4000 725000

28. HOME PRICES
area price
2600 550000
3000 565000
3200 610000
3600 680000
4000 725000
Given these home prices, find
out the price of homes whose
area is
3300 square feet
5000 square feet

29. SCATTER PLOT

30. BEST FIT LINE.

31. PREDICT HOME PRICES FOR A GIVEN AREA

32. PREDICT HOME PRICES FOR A GIVEN AREA (CONT.)

33. PREDICT HOME PRICES FOR A GIVEN AREA (CONT.)

34. SLOPE INTERSECTION EQUATION OF A STRAIGHT
LINE

35. PROGRAM IN PYTHON

36. EVALUATING MODEL PERFORMANCE
• The performance of a regression model can be understood by knowing the error
rate of the predictions made by the model. You can also measure the performance
by knowing how well your regression line fit the dataset.
• Let’s try to understand how to measure the performance of regression models.
• A good regression model is one where the difference between the actual or
observed values and predicted values for the selected model is small and unbiased
for train, validation and test data sets.

37. EVALUATING MODEL PERFORMANCE
• To measure the performance of your regression model, some statistical metrics are used. They
are-
• Mean Absolute Error(MAE)
• Root Mean Square Error(RMSE)
• Coefficient of determination or R2
• Adjusted R2

38. MEAN ABSOLUTE ERROR(MAE)
• This is the simplest of all the metrics. It is measured by taking the average of the absolute
difference between actual values and the predictions.

39. MEAN ABSOLUTE ERROR (MAE)

40. ROOT MEAN SQUARE ERROR(RMSE)
• The Root Mean Square Error is measured
by taking the square root of the average
of the squared difference between the
prediction and the actual value.
• It represents the sample standard
deviation of the differences between
predicted values and observed
values(also called residuals). It is
calculated using the following formula:

41. ROOT MEAN SQUARE ERROR(RMSE)

42. COEFFICIENT OF DETERMINATION OR R^2
• It measures how well the actual
outcomes are replicated by the
regression line.
• It helps you to understand how well the
independent variable adjusted with the
variance in your model.
• That means how good is your model
for a dataset.
• The mathematical representation for
R^2 is
Here, SSR = Sum Square of
Residuals(the squared difference
between the predicted and the
average value)
SST = Sum Square of Total(the
squared difference between the
actual and average value)

43. COEFFICIENT OF DETERMINATION OR R^2 (CONT.)
• Here the green line represents the regression line
and the red line represents the average line. The
differences in data points from these lines are
taken in the equation.
• Usually, the value of R^2 lies between 0 to 1(it
can be negative if the regression line somehow
has a worse fit than the average!). The closer its
value to one, the better your model is. This is
because either your regression line has well fitted
the dataset or the data points are distributed with
low variance. Which lessens the value of the Sum
of Residuals. Hence, the equation gets closer to
one.

44. THANK YOU