MACHINE LEARNING

Presenter: Dr. Amit Kumar Das

Professor,
Dept. of Computer Science and Engg.,
Institute of Engineering & Management.
WHAT IS LEARNING?
TYPES OF HUMAN LEARNING

 Learning through direct guidance from

expert – is just one form …

 Learning through indirect guidance

 Learning by self
WHAT IS MACHINE LEARNING?
WHAT IS MACHINE LEARNING?
TYPES OF MACHINE LEARNING
 Supervised learning – also called
predictive learning

 Unsupervised learning – also called

descriptive learning

 Reinforcement learning
 MACHINE LEARNING PROCESS
 What was the most difficult subject in the last
semester?

 What if, you had a list of all possible questions

with answers, and a photographic memory?
MACHINE LEARNING PROCESS
 Data Input – Past data or information is
utilized as a basis for future decision-making

 Abstraction – The input data is represented

in a broader way through the underlying
algorithm

 Generalization – The abstracted

representation is generalized to form a
framework for making decisions
TYPICAL ML PROBLEMS
 Prediction of results of a game
 Predicting whether a tumor is malignant or
benign
 Price prediction in domains like real estate,
stocks, etc.
 Demand forecasting in retails
 Customer segmentation
 Self-driven cars
PROBLEMS NOT TO BE CONSIDERED FOR ML

 Bank interest calculation

 Inventory management (except the demand

forecast module)

 Customer on-boarding (except risk prediction

module)

 Tasks in which humans are very effective or

frequent human intervention is needed. For
example, air traffic control
TYPES OF DATA
 Qualitative data (Categorical)
 Student Name, Blood group, Grade, etc.

 Quantitative data (Numerical)

 Temperature, Age, Weight, etc.
DATA EXPLORATION
 Understand the central tendency –
 Mean
 Median
 Mode

 Understand data spread

 Standard Deviation

 Understand data value position

DATA EXPLORATION – CENTRAL TENDENCY

Mean vs. Median for Auto MPG

DATA EXPLORATION – DATA SPREAD
 Consider the data values of two attributes
 Attribute 1 values – 44, 46, 48, 45 and 47
 Attribute 2 values – 34, 46, 59, 39 and 52
 Both the set of values have a mean and
median of 46.
 First set of values is more concentrated or
clustered around the mean / median value
DATA EXPLORATION – DATA VALUE POSITION

 Any data set attribute has five values

 Minimum
 First quartile (Q1)
 Median (Q2)
 Third quartile (Q3), and
 Maximum

Minimum Q1 Q3 Maximum

Median (Q2)
DATA EXPLORATION – BOX PLOT
DATA EXPLORATION – BOX PLOT
DATA QUALITY

 Most occurring data quality issues are:

 Missing values
 Outliers

Missing values of attribute “horsepower” in Auto MPG

 REMEDIATE DATA ISSUES
 Remove missing values / outliers – If
number of records are not many, remove them.
 Imputation - Impute the value with mean or
median or mode
 Capping - For values that lie outside the
1.5 X IQR limits, cap them by replacing the
observations below the lower limit with value of
5th percentile and those that lie above the upper
limit, with value of 95th percentile
 Estimate missing values – Assign attribute
values of similar data points in place of the
missing value
 ISSUES IN MACHINE LEARNING

 Relatively new and evolving technology

 In
different countries, rules and regulations,
cultural background, emotional maturity of
people are drastically different

 Biggestfears - potential breach of privacy,

discriminatory behaviour, resulting
discontent
WHAT IS MODELLING IN CONTEXT OF
MACHINE LEARNING?
WHAT ARE THE DIFFERENT ML
ALGORITHMS?

 Supervised
 Classification – KNN, Naive Bayes, Decision Tree, etc.

 Regression – Simple Linear Regression, Logistic

Regression

 Unsupervised
 Clustering – K-Means
 Market Basket Analysis
SUPERVISED LEARNING - CLASSIFICATION

Labelled Training Data

Classifier Classification Model

Test Data

Intel
SUPERVISED LEARNING - REGRESSION

y = α + βx
UNSUPERVISED LEARNING

Unlabelled Data

Unsupervised Learning Model

Grouped data / Clusters

UNSUPERVISED LEARNING - CLUSTERING

Cluster 2

Cluster 1

Cluster 3
Cluster 4
UNSUPERVISED LEARNING – MARKET BASKET
ANALYSIS
SELECTING A MODEL

 Predictive models (supervised)

 Predict the value of a category or class
 Problems that can be solved : Prediction of win/loss,
fraudulent transactions, etc.
 Examples : k-Nearest Neighbor (kNN), Naïve Bayes,
Decision Tree, etc.
 Predict numerical values of the target
 Problems that can be solved : Prediction of revenue
growth, rainfall amount, etc,
 Examples: Linear Regression, Logistic Regression, etc.
SELECTING A MODEL
 Descriptive
models
(unsupervised)
 Group together
similar data
instances
 Problems that can be
solved: Customer
grouping or
segmentation based
on social,
demographic, ethnic,
etc. factors
 Most popular model
for clustering is k-
Means
 TRAIN A MODEL – HOLDOUT METHOD
70% - 80% Training
Data

Input
Data Trained Model

Test
20% - 30% Data

Model Performance
K-FOLD CROSS-VALIDATION– OVERALL APPROACH
K-FOLD CROSS-VALIDATION– DETAILED APPROACH
K-FOLD CROSS-VALIDATION (CONTD.)
BOOTSTRAP SAMPLING / BOOTSTRAPPING
TRAIN A MODEL – UNDER VS. OVER FIT

Under fit Balanced fit Over fit

Under fit Balanced fit Over fit

TRAIN A MODEL – BIAS VS. VARIANCE
 EVALUATING A MODEL - CLASSIFICATION

Actual Outcome  True Positive (TP) –

Win Loss
Predicted win, Actual win
 True Negative (TN) –
Predicted loss, Actual loss
 False Positive (FP) –
Win

Predicted win, Actual loss

Predicted Outcome

True Positive (TP) False Positive (FP)  False Negative (FN) –

Predicted loss, Actual win

 For both TP and TN,

predicted outcome
Loss

matches actual
outcome. Hence, they
False Negative (FN) True Negative (TN)
are correct
classifications.
 EVALUATING A MODEL – CLASSIFICATION (CONTD.)
Actual
Actual Outcome
Actual Win Loss
Win Loss
Predicted Win 85 4
Predicted Loss 2 9
Win
Predicted Outcome

True Positive (TP) False Positive (FP)

Loss

False Negative (FN) True Negative (TN)

The percentage of misclassifications are indicated using error rate which is

measured as:

In context of the above confusion matrix,

 EVALUATING A MODEL – CLASSIFICATION (CONTD.)
where P(a) = proportion of observed agreement between actual
and predicted in overall data set =

P(pr) = proportion of expected agreement between actual and predicted data both in case
of class of interest as well as the other classes =

Note: Kappa value can be 1 at the maximum, which represents perfect agreement between model’s prediction and actual values.
EVALUATING A MODEL (ROC CURVE)
TPR =

FPR =

Receiver Operating Characteristic curve

 EVALUATING A MODEL (REGRESSION)
Value of the apartment unit 

Actual value

Error

Predicted value

Area (in square Feet) 

EVALUATING A MODEL (CLUSTERING)
“Clustering is in the eye of the beholder"

 Internal evaluation
 Silhouette width

 External evaluation
 Purity
EVALUATING A MODEL (CLUSTERING)
Cluster 2

Cluster 1
a(i)  Average distance between
ai2 ai1 the ith data instance and all other
data instances belonging to the
b14(1)
same cluster
ain_1 b(i)  Lowest average distance
b14(2)
between the i-the data instance and
b14(n4) data instances of all other clusters

Cluster 3
Cluster 4

Silhouette width calculation

ENSEMBLE
THANK YOU &
STAY TUNED!