UNIT-2

Machine Learning: Definition, Learning Methods: Supervised Learning,

Unsupervised Learning, Semi-Supervised Learning, Reinforcement Learning.

Q) Define Machine Learning. Explain types of ML Algorithms.

Machine learning is the study of computer algorithms that allow computer

programs to automatically improve through experience without being
explicitly programmed. Simply machine learning algorithms learn by
experience, similar to how humans. --Arthur Samuel(1959)

An algorithm can be thought of as a set of rules/instructions that a

computer programmer specifies which a computer can process.

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. -- Tom Mitchell(1997)

Eg. So if you want your program to predict, for example, traffic patterns at a
busy intersection (task T), you can run it through a machine learning
algorithm with data about past traffic patterns (experience E) and, if it has
successfully ―learned‖, it will then do better at predicting future traffic
patterns (performance measure P).

Need: Machine learning enables analysis of massive quantities of data. While

it generally delivers faster, more accurate results in order to identify
profitable opportunities or dangerous risks, it may also require additional
time and resources to train it properly. Combining machine learning with AI
and cognitive technologies can make it even more effective in processing
large volumes of information.

Machine learning algorithms are often categorized as supervised or

unsupervised.

1. Supervised machine learning algorithms can apply what has been

learned in the past to new data using labeled examples to predict
future events.

Starting from the analysis of a known training dataset, the learning

algorithm produces an inferred function to make predictions about the
output values.
The system is able to provide targets for any new input after sufficient
training.
The learning algorithm can also compare its output with the correct,
intended output and find errors in order to modify the model accordingly.

Eg. suppose you are given an basket filled with different kinds of fruits.

Now the first step is to train the machine with all different fruits one by one
like this:
If shape of object is rounded and depression at top having color Red
then it will be labelled as –Apple.
If shape of object is long curving cylinder having color Green-Yellow
then it will be labelled as –Banana.

Now suppose after training the data, you have given a new separate fruit say
Banana from basket and asked to identify it.
Since machine has already learned the things from previous data and this
time have to use it wisely. It will first classify the fruit with its shape and
color, and would confirm the fruit name as BANANA and put it in Banana
category. Thus machine learns the things from training data(basket
containing fruits) and then apply the knowledge to test data(new fruit).

Supervised learning classified into two categories of algorithms:

Classification: A classification problem is when the output variable is a
category, such as ―Red or blue or disease and no disease.

Regression: A regression problem is when the output variable is a real

value, such as price or weight or height.
 2. In contrast, unsupervised machine learning algorithms are used when
the information used to train is neither classified nor labeled.

Unsupervised learning studies how systems can infer a function to describe

a hidden structure from unlabeled data.

The system doesn‘t figure out the right output, but it explores the data and
can draw inferences from datasets to describe hidden structures from
unlabeled data.

Eg.

suppose it is given an image having both dogs and cats which have not seen
ever. Thus machine has no any idea about the features of dogs and cat so
we can‘t categorize it in dogs and cats.

But it can categorize them according to their similarities, patterns and

differences i.e., we can easily categorize the above picture into two parts.
First may contain all pictures having dogs in it and second part may contain
all pictures having cats in it based on some properties like color, size etc.
Here you didn‘t learn anything before, means no labeled data.

3. Semi-supervised machine learning algorithms fall somewhere in

between supervised and unsupervised learning, since they use both
labeled and unlabeled data for training – typically a small amount of
labeled data and a large amount of unlabeled data.

The systems that use this method are able to considerably improve learning
accuracy.

Usually, semi-supervised learning is chosen when the acquired labeled data

requires skilled and relevant resources in order to train it / learn from it.
Otherwise, acquiring unlabeled data generally doesn‘t require additional
resources.

4. Reinforcement learning is a type of machine learning in which a

computer learns to perform a task through repeated trial-and-error
interactions with a dynamic environment. This learning approach
enables the computer to make a series of decisions that maximize a
reward metric for the task without human intervention and without
being explicitly programmed to achieve the task.
Simple reward feedback is required for the agent to learn which action is best
this is known as the reinforcement signal.

Eg.

Fig. RL in dog training

The goal of reinforcement learning in this case is to train the dog (agent) to
complete a task within an environment, which includes the surroundings of
the dog as well as the trainer.

First, the trainer issues a command, which the dog observes (observation).
The dog then responds by taking an action. If the action is close to the
desired behavior, the trainer will likely provide a reward, such as a food
treat; otherwise, no reward or a negative reward will be provided.

At the beginning of training, the dog will likely take more random actions
like rolling over when the command given is ―sit,‖ as it is trying to associate
specific observations with actions and rewards. This association, or
mapping, between observations and actions is called policy.

From the dog‘s perspective, the ideal case would be one in which it would
respond correctly to every command, so that it gets as many treats as
possible.
So, the whole meaning of reinforcement learning training is to ―tune‖ the
dog‘s policy so that it learns the desired behaviors that will maximize some
reward. After training is complete, the dog should be able to observe the
owner and take the appropriate action, for example, sitting when
commanded to ―sit‖ by using the internal policy it has developed.
Q) Differentiate ML Vs Classical/Traditional Algorithms.

Traditional Programming
Traditional programming is a manual process—meaning a programmer
creates the program.

Machine Learning
Unlike traditional programming, machine learning is an automated
process. ML algorithm automatically formulates the rules from the data.

ML vs Classical Algorithms:

 ML algorithms do not depend on rules defined by human experts.

Instead, they process data in raw form like text, emails, documents,
social media content, images, voice and video.
 An ML system is truly a learning system if it is not programmed to
perform a task, but is programmed to learn to perform the task
 Most ML models are uninterpretable, and for these reasons they are
usually unsuitable when the purpose is to understand relationships.
The mostly work well where one only needs predictions.
 One of the key differences is that classical approaches have a more
rigorous mathematical approach while machine learning algorithms
are more data-intensive

Q) Define a Data Object.

Data sets are made up of data objects. A data object represents an entity in
a database table. Data objects are typically described by attributes. Data
objects can also be referred to as samples, examples, instances, data points,
or objects. If the data objects are stored in a database, they are data tuples.
That is, the rows of a database correspond to the data objects, and the
columns correspond to the attributes.

Q) Define an attribure. Explain types of attributes with an example.

An attribute is a data field, representing a characteristic or feature of a data

object. It also called dimension, feature, variable. The type of an attribute is
determined by the set of possible values—nominal, binary, ordinal, or
numeric.
Nominal attributes: Each value represents some kind of category, code, or
state, and so nominal attributes are also referred to as categorical. The
values do not have any meaningful order.
Eg. marital status, occupation, ID numbers, zip codes, hair color.

Ordinal Attributes: An ordinal attribute is an attribute with possible values

that have a meaningful order or ranking among them, but the magnitude
between successive values is not known.
Eg. Grades, Qualification.

Binary Attributes: A binary attribute is a nominal attribute with only two

categories or states: 0 or 1, where 0 typically means that the attribute is
absent, and 1 means that it is present.
Binary attributes are referred to as Boolean if the two states correspond to
true and false.

Symmetric binary: A binary attribute is symmetric if both of its

states are equally valuable and carry the same weight; that is, there is no
preference on which outcome should be coded as 0 or 1.
Eg. gender
Asymmetric binary: A binary attribute is asymmetric if the outcomes
of the states are not equally important, such as the positive and negative
outcomes of a medical test
Eg. test for fever, result of an examination.

Numeric Attributes: A numeric attribute is quantitative; that is, it is a

measurable quantity, represented in integer or real values. Numeric
attributes can be interval-scaled or ratio-scaled.

Interval-scaled attributes are measured on a scale of equal-size

units. The values of interval-scaled attributes have order and can be
positive, 0, or negative. It is not a Zero-Point.
E.g. temperature in C˚or F˚, calendar dates

A ratio-scaled attribute is a numeric attribute with an inherent zero-

point. That is, if a measurement is ratio-scaled, we can speak of a value as
being a multiple (or ratio) of another value.
Eg. temperature in Kelvin, length, counts, monetary quantities

Discrete versus Continuous Attributes

Discrete Attribute: A discrete attribute has a finite or countably infinite

set of values, which may or may not be represented as integers. Binary
attributes are a special case of discrete attributes.
E.g. zip codes, profession, or the set of words in a collection of documents.
Continuous Attribute: If an attribute is not discrete, it is continuous. It
has real numbers as attribute values. Continuous attributes are typically
represented as floating-point variables.
E.g. temperature, height, or weight of a person.

Q) Explain indetail about Supervised Learning approach.

Predictive or Supervised learning approach, the goal is to learn a mapping

from inputs x to outputs y, given a labeled set of input-output pairs.
D = {(xi, yi)}i=1to N.
Here, D is called the training set, and N is the number of training examples.

Classification:
Classification is a type of supervised learning. It specifies the class to which
data elements belong to and is best used when the output has finite and
discrete values. It predicts a class for an input variable as well. Here the
class label is categorical.

Fig. Illustration of Classification for Loan Sanction.

Each training input xi is a D-dimensional vector of numbers, representing,

say, the age, ed, employ, address, income, debtinc, creddebt and othdebt of a
person. These are called features, attributes, dimensions, predictor variable,
independent variables, columns or covariates. In general, however, xi could
be a complex structured object, such as an image, a sentence, an email
message, a time series, a molecular shape, a graph, etc

Similarly the form of the output, dependent variable or response variable can
in principle is anything, but most methods assume that yi is a categorical or
nominal variable from some finite set, yi ∈ {1, . . . , C}, where C is no.of
classes.

If C = 2, this is called binary classification (Eg. 0 or 1, yes/no)

if C > 2, this is called multiclass classification.(Eg. {0,1 or 2},
{high,low,medium})
When yi is categorical, the problem is known as classification or pattern
recognition, and when yi is real-valued, the problem is known as regression.

We assume y = f(x) for some unknown function f, and the goal of learning is
to estimate the function f given a labelled training set, and then to make
predictions using ˆy = ˆ f(x). (We use the hat symbol to denote an estimate.)

Our main goal is to make predictions on new inputs, meaning ones that we
have not seen before called generalization. Since predicting the response on
the training set is easy as we can just look up the answer.

Need for probabilistic predictions:

In above fig., the yellow circle is harder to classify, since some yellow things
are labeled y = 1 and some are labeled y = 0, and some circles are labeled y =
1 and some y = 0.

Consequently it is not clear what the right label should be in the case of the
yellow circle. Similarly, the correct label for the blue arrow is unclear.

To handle ambiguous cases, such as the yellow circle above, it is desirable to

return a probability.

Given a probabilistic output, we can always compute our ―best guess‖ as to

the ―true label‖ using

This corresponds to the most probable class label, and is called the mode of
the distribution p(y|x,D); it is also known as a MAP estimate (MAP stands for
maximum a posteriori).
Real-world applications of classification:
Classification is probably the most widely used form of machine learning,
and has been used to solve many interesting and often difficult real-world
problems.

1. In document classification, the goal is to classify a document, such

as a web page or email message, into one of C classes, that is, to
compute p(y = c|x,D), where x is some representation of the text.
A special case of this is email spam filtering, where the classes are
spam y = 1 or ham y = 0.
2. Image classification:
i. Classifying flowers: The goal is to learn to distinguish three
different kinds of iris flower, called setosa, versicolor and
virginica.

ii. Hand writing recognition

MNIST, which stands for ―Modified National Institute of Standards, This

dataset contains 60,000 training images and 10,000 test images of the digits
0 to 9, as written by various people. The images are size 28×28 and have
grayscale values in the range 0 : 255.
 3. Object detection and recognition: A harder problem is to find objects
within an image; this is called object detection or object localization. A
special case of this is face detection.

Fig. Example for Object Detection

4. In banking, predicting loan sanction.

5. Speech Identification

Regression:

Regression is just like classification except the response variable is

continuous.
Regression analysis consists of a set of machine learning methods that allow
us to predict a continuous outcome variable (y) based on the value of one or
multiple predictor variables (x).

In linear regression, the data are modeled to fit a straight line.

y = wx+b, variable, y (called a response variable), can be
modeled as a linear function of another random variable, x (called a
predictor variable) where the variance of y is assumed to be constant.

In the context of data mining, x and y are numeric database attributes. The
coefficients, w and b (called regression coefficients), specify the slope of the
line and the y-intercept, respectively.
Eg.

Fig. Linear Regression

In non-linear regression, the data can‘t be modelled to fit a straight line i.e.
modelled to fit a curve as shown in below figure.

For example the equation can have the form: XTheta

Here are some examples of real-world regression problems.

• Predict tomorrow‘s stock market price given current market
conditions and other possible side information.
• Predict the age of a viewer watching a given video on YouTube.
• Predict the location in 3d space of a robot arm end effector, given
control signals (torques) sent to its various motors.
• Predict the amount of prostate specific antigen (PSA) in the body as a
function of a number of different clinical measurements.
• Predict the temperature at any location inside a building using
weather data, time, door sensors, etc.

Q) Differentiate Simple linear regression and Multi linear regression.

Linear regression models the relationship between a dependent variable and

one or more explanatory variables using a linear function can be formulated
as below:
y = bo + b1*x1
Eg. Prediction of height based on age.
Feature(X): age
Class(y): height
Regression equation will be: height = w*age + b

If two or more explanatory variables have a linear relationship with the

dependent variable, the regression is called a multiple linear regression,
which can be formulated as below:
y = bo + b1*x1 + b2*x2 .. + bn*xn

where,
y is the response variable.
 a, b1, b2...bn are the coefficients.
x1, x2, ...xn are the predictor variables.
Eg. Prediction of height based on age and gender.
Features(X):age, gender
class(y): height
Regression equation will be: height = w1*age + w2*gender + b

Q) Explain about KNN classification

Eager learners, when given a set of training tuples, will construct a
generalization (i.e., classification) model before receiving new (e.g., test)
tuples to classify.

Eg. decision tree induction, Bayesian classification, rule-based

classification, classification by backpropagation, support vector machines,
and classification based on association rule mining.
A lazy learner simply stores the given set of training tuples and only when
it sees the test tuple does it perform generalization to classify the tuple
based on its similarity to the stored training tuples.

Eg. KNN
k-Nearest-Neighbor Classifiers

Algorithm:

◦ Initialize k value.
◦ Calculate the distance between test data and each row of training
data. Here we will use Euclidean distance as our distance metric since it‗s
the most popular method. The other metrics that can be used are
manhatten, cosine, etc.
◦ Sort the calculated distances in ascending order based on distance
values
◦ Get top k rows from the sorted array
◦ Get the most frequent class of these rows
◦ Return the predicted class

Eg.

Fig. KNN Classifier

 In the above fig. the new data point star belongs to circle class as 3 nearest
neighbours are circles.

Advantages:
 Quick calculation time
 Simple algorithm – to interpret
 useful for regression and classification
 No assumptions about data – no need to make additional
assumptions, tune several parameters, or build a model. This
makes it crucial in nonlinear data case.
Disadvantages:
 Accuracy depends on the quality of the data
 With large data, the prediction stage might be slow
 Require high memory – need to store all of the training data

Q) Briefly explain about logistic regression.

Logistic regression is a supervised learning classification algorithm used to

predict the probability of a target variable. The nature of target or dependent
variable is dichotomous, which means there would be only two possible
classes.

Types of Logistic Regression

Generally, logistic regression means binary logistic regression having binary

target variables, but there can be two more categories of target variables
that can be predicted by it. Based on the number of categories, Logistic
regression can be divided into following types:
 Binary Logistic Regression Model: The simplest form of logistic
regression is binary or binomial logistic regression in which the target
or dependent variable can have only 2 possible types either 1 or 0.

Eg. Spam detection is a binary classification problem where we are given

an email and we need to classify whether or not it is spam. If the email is
spam, we label it 1; if it is not spam, we label it 0. In order to apply
Logistic Regression to the spam detection problem, the following features
of the email are extracted:

Sender of the email

Number of typos in the email
Occurrence of words/phrases like ―offer‖, ―prize‖, ―free gift‖, etc.
The resulting feature vector is then used to train a Logistic classifier
which emits a score in the range 0 to 1. If the score is more than 0.5, we
label the email as spam. Otherwise, we don‘t label it as spam.
 Multinomial Logistic Regression Model: Another useful form of
logistic regression is multinomial logistic regression in which the target or
dependent variable can have 3 or more possible unordered types i.e. the
types having no quantitative significance.

Eg. Entering high school students make program choices among general
program, vocational program and academic program. Their choice might be
modeled using their writing score and their social economic status.

Q) Differentiate Key Differences between Linear and Logistic

Regression.

1. The Linear regression models data using continuous numeric value. As

against, logistic regression models the data in the binary values.
2. Linear regression requires establishing the linear relationship among
dependent and independent variable whereas it is not necessary for
logistic regression.
3. In the linear regression, the independent variable can be correlated
with each other. In contrast the logistic regression, the variable must not
be correlated with each other.

Q) Explain indetail about Unsupervised learning.

In unsupervised learning, we are just given output data, without any inputs.
The goal is to discover ―interesting structure‖ in the data; this is sometimes
called knowledge discovery.

Unlike supervised learning, we are not told what the desired output is for
each input. Instead, we will formalize our task as one of density estimation,
that is, we want to build models of the form p(xi|θ).

Difference between Supervised and Unsupervised learning:

1. Supervised learning is conditional density estimation, whereas
unsupervised learning is unconditional density estimation.
2. xi is a vector of features, so we need to create multivariate probability
models. By contrast, in supervised learning, yi is usually just a single
variable that we are trying to predict.
Unsupervised learning is arguably more typical of human and animal
learning. It is also more widely applicable than supervised learning, since it
does not require a human expert to manually label the data. Labeled data is
not only expensive to acquire, but it also contains relatively little information,
certainly not enough to reliably estimate the parameters of complex models.

Here are some real world applications of clustering:

• In astronomy, the autoclass system (Cheeseman et al. 1988) discovered a
new type of star, based on clustering astrophysical measurements.
• In e-commerce, it is common to cluster users into groups, based on their
purchasing or web-surfing behavior, and then to send customized targeted
advertising to each group.
• In biology, it is common to cluster flow-cytometry data into groups, to
discover different sub-populations of cells.
• In banking, it is used to find credit card fraud detection by using user
transaction behavior.

Canonical examples of unsupervised learning:

1. Discovering clusters
consider the problem of clustering data into groups.

The above figure, plots some 2d data, representing the height and weight of
a group of 210 people. It seems that there might be various clusters, or
subgroups, although it is not clear how many. Let K denote the number of
clusters. Our first goal is to estimate the distribution over the number of
clusters, p(K|D); this tells us if there are subpopulations within the data. For
simplicity, we often approximate the distribution p(K|D) by its mode,
K∗ = arg maxK p(K|D).

In the supervised case, we were told that there are two classes (male and
female), but in the unsupervised case, we are free to choose as many or few
clusters as we like. Picking a model of the “right” complexity is called
model selection.

Our second goal is to estimate which cluster each point belongs to. Let zi(zi
is an example of a hidden or latent variable, since it is never observed in the
training set.) ∈ {1, . . . , K} represent the cluster to which data point i is
assigned.

2. Discovering latent factors

When dealing with high dimensional data, it is often useful to reduce the
dimensionality by projecting the data to a lower dimensional subspace which
captures the ―essence‖ of the data. This is called dimensionality reduction.
Such low dimensional representations often result in better predictive
accuracy, because they focus on the ―essence‖ of the object, filtering out
inessential features.
Also, low dimensional representations are useful for enabling fast nearest
neighbor searches and two dimensional projections are very useful for
visualizing high dimensional data.

The most common approach to dimensionality reduction is called principal

components analysis (PCA).

PCA in particular, has been applied in many different areas. Some examples
include the following:
• In biology, it is common to use PCA to interpret gene microarray data, to
account for the fact that each measurement is usually the result of many
genes which are correlated in their behavior by the fact that they belong to
different biological pathways.
• In natural language processing, it is common to use a variant of PCA called
latent semantic analysis for document retrieval.
• In signal processing (e.g., of acoustic or neural signals), it is common to use
ICA (which is a variant of PCA) to separate signals into their different sources
• In computer graphics, it is common to project motion capture data to a low
dimensional space, and use it to create animations.

3. Discovering graph structure

Sometimes we measure a set of correlated variables, and we would like to

discover which ones are most correlated with which others. This can be
represented by a graph G, in which nodes represent variables, and edges
represent direct dependence between variables. We can then learn this graph
structure from data, i.e., we compute ˆG = argmax p(G|D).

Fig. sparse undirected Gaussian graphical model learned using graphical

lasso applied to some flow cytometry data which measures the
phosphorylation status of 11 proteins.

As with unsupervised learning in general, there are two main applications for
learning sparse graphs: to discover new knowledge, and to get better joint
probability density estimators.
 4. Matrix Completion

Sometimes we have missing data, that is, variables whose values are
unknown.
For example, we might have conducted a survey, and some people might not
have answered certain questions. Or we might have various sensors, some of
which fail.
The corresponding design matrix will then have ―holes‖ in it; these missing
entries are often represented by NaN, which stands for ―not a number‖. The
goal of imputation is to infer plausible values for the missing entries also
called matrix completion.

Applications of Matrix completion:

1. Image inpainting:
The goal is to “fill in” holes (e.g., due to scratches or occlusions) in an
image with realistic texture, where we denoise the image, as well as impute
the pixels hidden behind the occlusion. This can be tackled by building a
joint probability model of the pixels, given a set of clean images, and then
inferring the unknown variables (pixels) given the known variables(pixels).

Fig. (a) A noisy image with an occluder. (b) An estimate of the underlying
pixel intensities, based on a pairwise MRF model.

2. Collaborative filtering:

Collaborative filtering is the predictive process behind recommendation

engines. Recommendation engines analyze information about users with
similar tastes to assess the probability that a target individual will enjoy
something, such as a video, a book or a product. Collaborative filtering is
also known as social filtering.

Eg. Predicting which movies people will want to watch based on how they,
and other people, have rated movies which they have already seen.
Fig. movie-rating data. Training data is in red, test data is denoted by ?,
empty cells are unknown.

3. Market basket analysis

A set of items is referred to as an itemset. The occurrence or frequency of
an itemset in the number of transactions that contain the itemset is called
count of the itemset.

If the count of the itemset satisfy a minimum support threshold then the
itemset is said to be a frequent itemset.
Frequent itemset mining leads to the discovery of associations and
correlations among items in large transactional or relational data sets.
A typical example of frequent itemset mining is market basket analysis.
This process analyzes customer buying habits by finding associations
between the different items that customers place in their ―shopping
baskets‖.

Eg. 1. Bread => Jam and 2. Bread, Milk => Sugar

Fig. Market basket analysis.

The discovery of these associations can help retailers develop marketing

strategies by gaining insight into which items are frequently purchased
together by customers. For instance, if customers are buying milk, how
likely are they to also buy bread.
Q) Explain about Partitioning cluster algorithms with an example.

Given a set of n objects, a partitioning method constructs k partitions of the

data, where each partition represents a cluster and k <= n.

The basic partitioning methods typically adopt exclusive cluster

separation. That is, each object must belong to exactly one group.
Given k, the k-means algorithm consists of four steps:

1. Select initial centroids at random.

2. Assign each object to the cluster with the nearest centroid.
3. Compute each centroid as the mean of the objects assigned to it.
4. Repeat previous 2 steps until no change.
Euclidean distance:

Updating centroid point:

Algorithm: The k-means algorithm for partitioning, where each cluster's center is
represented by the mean value of the objects in the cluster.
Input:
k: the number of clusters,
D: a data set containing i objects.
Output: A set of k clusters.
Method:
(1) arbitrarily choose k objects from D as the initial cluster centers;
(2) repeat
(3) (re)assign each object to the cluster to which the object is the most similar,
based on the mean value of the objects in the cluster;
(4) update the cluster means, that is, calculate the mean value of the objects for
each cluster;
(5) until no change;
Eg. Consider the following 2-dimensional data objects:

Sl.No. A B

1 1 1

2 1.5 2

3 3 4

4 5 7

5 3.5 5

6 4.5 5

7 3.5 4.5

Let us start with random k=2 and the centroids be c1= (1,1) and c2= (5,7)

Now we find the Euclidean distance from all the given points to the centroids taken and
is as shown below:

st
From the above table it is clear that the points (1,1), (1.5,2) and (3,4) falls in the 1 cluster
nd
and remaining points fall in the 2 cluster.

So we find the new centroids as:

C1 = [(1+1.5+3)/3, (1+2+4)/3] = (1.83, 2.33)
C2 = [(5+3+4.5+3.5)/4, (7+5+5+4.5)/4] = (4, 5.37)
The distances from all the given points to the updated centroids are as follows:

st
From the above table it is clear that the points (1,1) and (1.5,2) fall in the 1 cluster
nd
and the remaining into the 2 cluster.

Since we get different points into the cluster compared to the previous step we continue
the process and update the centroids as follows:
C1 = [(1+1.5)/2, (1+2)/2] = (1.25, 1.5)
C2 = [(3+5+3+4.5+3.5)/5, (4+7+5+5+4.5)/5] = (4.8, 5.1)
Now again we calculate the distance between all the given points and the updated C1 and
C2 as follows.

st nd
From the above table it is evident that the points in the 1 cluster and 2 cluster are same
as the previous step.
So we stop the iterations and the final cluster of points are cluster1 = {(1,1) , (1.5,2) }
and cluster2 = {(3,4),(5,7),(3,5),(4.5,5),(3.5,4.5)}
Strengths
–Relatively efficient: O(tkn), where n is # objects, k is # clusters, and t is
# iterations. Normally, k, t << n.
–Often terminates at a local optimum. The global optimum may be found
using techniques such as simulated annealing and genetic algorithms
Weaknesses
–Applicable only when mean is defined.
–Need to specify k, the number of clusters, in advance
–Trouble with noisy data and outliers
–Not suitable to discover clusters with non-convex shapes.

Q) Explain about Parametric vs non-parametric models in Machine

Learning.

A learning model that summarizes data with a set of parameters of fixed size
(independent of the number of training examples) is called a parametric
model. No matter how much data you throw at a parametric model, it won‘t
change its mind about how many parameters it needs.

The algorithms involve two steps:

1. Select a form for the function.

2. Learn the coefficients for the function from the training data.
An easy to understand functional form for the mapping function is a line, as
is used in linear regression:

b0 + b1*x1 + b2*x2 = 0
Where b0, b1 and b2 are the coefficients of the line that control the intercept
and slope, and x1 and x2 are two input variables.
Some more examples of popular nonparametric machine learning algorithms
are:
 Naive Bayes
 Logistic Regression

Benefits of Parametric Machine Learning Algorithms:

 Simpler: These methods are easier to understand and interpret

results.
 Speed: Parametric models are very fast to learn from data.
  Less Data: They do not require as much training data and can work
well even if the fit to the data is not perfect.
Limitations of Parametric Machine Learning Algorithms:

 Constrained: By choosing a functional form these methods are highly

constrained to the specified form.
 Limited Complexity: The methods are more suited to simpler
problems.
 Poor Fit: In practice the methods are unlikely to match the underlying
mapping function.

Nonparametric methods seek to best fit the training data in constructing

the mapping function, while maintaining some ability to generalize to
unseen data. As such, they are able to fit a large number of functional
forms.
An easy to understand nonparametric model is the k-nearest neighbors
algorithm that makes predictions based on the k most similar training
patterns for a new data instance. The method does not assume anything
about the form of the mapping function other than patterns that are close
likely to have a similar output variable.
Some popular nonparametric machine learning algorithms are:
 k-Nearest Neighbors
 Decision Trees like CART and C4.5
 Support Vector Machines

Benefits of Nonparametric Machine Learning Algorithms:

 Flexibility: Capable of fitting a large number of functional forms.

 Power: No assumptions (or weak assumptions) about the underlying
function.
 Performance: Can result in higher performance models for prediction.
Limitations of Nonparametric Machine Learning Algorithms:

 More data: Require a lot more training data to estimate the mapping
function.
 Slower: A lot slower to train as they often have far more parameters to
train.
 Overfitting: More of a risk to overfit the training data and it
is harder to explain why specific predictions are made.
Q) Briefly explain Semi supervised learning.
Semi-supervised machine learning is a combination
of supervised and unsupervised machine learning methods.
With more common supervised machine learning methods,
you train a machine learning algorithm on a ―labeled‖ dataset in which each
record includes the outcome information.
This allows the algorithm to deduce patterns and identify relationships
between your target variable and the rest of the dataset based on information
it already has.
In contrast, unsupervised machine learning algorithms learn from a dataset
without the outcome variable.
In semi-supervised learning, an algorithm learns from a dataset that
includes both labeled and unlabeled data, usually mostly unlabeled.
Need: When you don‘t have enough labeled data to produce an accurate
model and you don‘t have the ability or resources to get more data, you can
use semi-supervised techniques to increase the size of your training data.

Label Propagation Algorithm (LPA) is an iterative algorithm where we

assign labels to unlabelled points by propagating labels through the dataset.
Eg.
Assume that we have a network of people as given below with two label
classes ―interested in cricket‖ and ―not interested in cricket‖. So the question
is, can we predict whether the remaining people are interested in cricket or
not?
 For LPA to work in this case, we have to make an assumption; an edge
connecting two nodes carry a notion of similarity. i.e. if two people are
connected together, that means that it is highly likely that these two people
share the same interests.

Consider the sample graph given in below, where we have 2 label classes (red and green)
and 4 nodes coloured (2 for each class). We want to predict the label of node 4.

Fig . Sample graph

We can walk randomly in the graph, starting from node 4 until we meet any labelled
node. When we hit a labelled node, we stop the walk. Hence, these labelled nodes are
known as absorbing states. Let’s consider all the possible walks from node 4. Out of all
the possible walk, the following walks will end in a green node.
1. 4 → 9 → 15 → 16
2. 4 → 9 → 13 → 14
3. 4 → 9 → 13 → 15 → 16
4. 4 → 9 → 15 → 13 → 14
The following walks will end in a red node.
1. 4→7→8
2. 4→7→6→5→1
3. 4→5→1
4. 4→5→6→7→8
5. 4→2→1
Based on all the possible random walks starting from node 4, we can see
that the majority of the walks end in a red node. So, we can colour node 4 in
red. This is the basic intuition behind LPA.
Q) What is Reinforcement Learning? Explain about agent in RL.

Reinforcement learning (RL) is an area of machine learning concerned with

how software agent have to take actions in an environment in order to
maximize the notion of cumulative reward.

Fig. Illustrating of RL

Agent
Agents are the software programs that make intelligent decisions and they
are basically learners in RL.
Agents take action by interacting with the environment and they receive
rewards based on their actions.
Eg. Mario navigating in a video game.

Eg. 1: A human agent has eyes, ears, and other organs for sensors and
hands, legs, vocal tract, and so on for actuators.
Eg. 2: A robotic agent might have cameras and infrared range finders for
sensors and various motors for actuators.
Eg. 3: A software agent receives keystrokes, file contents, and network
packets as sensory inputs and acts on the environment by displaying on the
screen, writing files, and sending network packets.
Eg. 4: Vaccum cleaner that cleans blocks A & B. A reward of +1 will be given
if it sucks dust and a reward of 0 will be given if in same block and -1 if
moves from one block to another.
The elements of RL are:

Fig. Elements of RL

Policy function
A policy defines the agent's behavior in an environment. The way in which
the agent decides which action to perform depends on the policy.
A policy is often denoted by the symbol 𝛑. A policy can be in the form of a
lookup table or a complex search process.

Eg. If you want to reach your office from home; there will be different routes
to reach your office, and some routes are shortcuts, while some routes are
long. These routes are called policies because they represent the way in
which we choose to perform an action to reach our goal.
Value function
 A value function denotes how good it is for an agent to be in a
particular state. It is dependent on the policy and is often denoted by
v(s). It is equal to the total expected reward received by the agent
starting from the initial state.
 There can be several value functions. The optimal value function is the
one that has the highest value for all the states compared to other
value functions.
 Similarly, an optimal policy is the one that has the optimal value
function.

Model
Model is the agent's representation of an environment. The learning can be
of two types—
In model-based learning, the agent exploits previously learned
information to accomplish a task. Eg. Epsilon-Greedy approach, random
selection approach.
whereas in model-free learning, the agent simply relies on a trial-and-error
experience for performing the right action. Eg. Q-learning or policy gradient.

Eg. If you want to reach your office from home faster. In model-based
learning, you simply use a previously learned experience (map) to reach the
office faster, whereas in model-free learning you will not use a previous
experience and will try all different routes and choose the faster one.

Reward Function Engineering determines the rewards for actions.

Q) Explain the Nature of Environment in RL.

Task environments, which are essentially the ―problems‖ to which rational

agents are the ―solutions.‖ we had to specify the performance measure, the
environment, and the agent‘s actuators and sensors. We group all these
under the heading of the task environment.
For the acronymically minded, we call this as the PEAS
(Performance, Environment, Actuators, Sensors) description. In designing
an agent, the first step must always be to specify the task environment as
fully as possible.
Eg.: an automated taxi driver.
The full driving task is extremely open-ended. There is no limit to the novel
combinations of circumstances that can arise.
Type Performance Environment Actuators Sensors
Agent Measure
Type
Taxi Safe, fast, Roads, other Steering, Cameras, sonar,
driver legal, traffic, accelerator, speedometer, GPS,
comfortable pedestrians, brake, odometer,
trip, customers signal, accelerometer, engine
maximize horn, sensors,
profits display keyboard.
Q) Explain about different types of Enviornment in RL.

Fully observable vs. partially Fully observable:

If an agent‘s sensors give it access to the complete state of the
environment at each point in time, then we says that the task environment
is fully observable.
An environment might be partially observable because of noisy and
inaccurate sensors or because parts of the state are simply missing from the
sensor data.
Eg. a vacuum agent with only a local dirt sensor cannot tell whether there is
dirt in other squares.

Single agent Vs. Multi agent:

An environment in which only single agent exists is called a Single
Agent Environment.
Eg. Crossword puzzle.
An environment in which more than one single agent exists is called a
Multi-agent Environment.
 Eg. Chess is a competitive multiagent environment because both the agents
try to maximize their performance (win). In the taxi-driving environment, on
the other hand, avoiding collisions maximizes the performance measure of
all agents, so it is a partially cooperative multiagent environment.

Deterministic Vs. Stochastic:

If the next state of the environment is completely determined by the current
state and the action executed by the agent, then we say the environment is
deterministic; otherwise, it is stochastic. If the environment is partially
observable, however, then it could appear to be stochastic.
Eg. The vacuum world as we described it is deterministic, but variations
can include stochastic elements such as randomly appearing dirt and an
unreliable suction mechanism.
Taxi driving is clearly stochastic in this sense, because one can never
predict the behavior of traffic exactly. We say an environment is uncertain if
it is not fully observable or not deterministic.

Episodic Vs. Sequential:

In an episodic task environment, the agent‘s experience is divided
into atomic episodes. In each episode the agent receives a percept and then
performs a single action. Crucially, the next episode does not depend on the
actions taken in previous episodes. Many classification tasks are episodic.
Episodic environments are much simpler than sequential environments
because the agent does not need to think ahead.
Eg. an agent that has to spot defective parts on an assembly line bases each
decision on the current part, regardless of previous decisions; moreover, the
current decision doesn‘t affect whether the next part is defective.

In sequential environments the current decision could affect all future

decisions. Eg. Chess and taxi driving.

Static vs. Dynamic:

If the environment can change while an agent is deliberating, then we say
the environment is dynamic for that agent; otherwise, it is static.
Eg. Taxi Driving is dynamic
Static environments are easy to deal with because the agent need not keep
looking at the world while it is deciding on an action, nor need it worry
about the passage of time.
Eg. Crossword puzzle is static

If the environment itself does not change with the passage of time but the
agent‘s performance score does, then we say the environment is
semidynamic.
Eg. Chess, when played with a clock, is semidynamic.

Discrete vs. Continuous:

The discrete/continuous distinction applies to the state of the environment,
to the way time is handled, and to the percepts and actions of the agent.
Eg. Input from digital cameras is discrete; Taxi driving is a continuous-state
Known vs. Unknown:
This distinction refers not to the environment itself but to the agent‘s state
of knowledge about the environment. In a known environment the outcomes
(or outcome probabilities if the environment is stochastic) for all actions are
given. Obviously, if the environment is unknown, the agent will have to learn
how it works in order to make good decisions.

Q) Differentiate Supervised, Unsupervised and Reinforcement Learnings.

Criteria Unsupervised ML Reinforcement ML
Supervised ML

Definition Learns by using Trained using Works on interacting

labelled data unlabelled data with the environment
without any guidance.

Type of data Labelled data Unlabelled data No - predefined data

Type of problems Regression and Association and Exploitation or

classification Clustering Exploration

Supervision Extra supervision No supervision No supervision

Algorithms Linear Regression, K - Means, Q - Learning, SARSA,

Logistic Regression, Hierarchical etc.
SVM, KNN etc. clustering,
Apriori, etc.

Aim Calculate outcomes Discover underlying Learn a series of

patterns action
Application Risk Evaluation, Recommendation Self Driving Cars,
Forecast Sales System, Anomaly Gaming, Healthcare
Detection

Q) Fit a straight line to the following data

x 0 1 2 3 4
y 1 1.8 3.3 4.5 6.3

Sol:
Let the required linear equation be

y=a1x+a0. We know,

Here m = 5

There fore, we get equations as

16.9=5a0+10a1
47.1=10a0+30a1
By solving above two equations we get a0=0.72 a1=1.33
Therefore, required equation is y = 1.33x + 0.72.

bx
Q) Fit a curve of the type y=ae to the following data

x 0 1 2 3

y 1.05 2.1 3.85 8.3

The given relation is y=aebx
Taking logarithms on both sides we obtain log y = log a + bx----(1)

Let log y = Y and x = X, log a = a0 and b = a1

The eq(1) takes the form Y = a0 + a1X, which is a straight line.

We know,

And m=4

4.255 = 4a0 + 6a1 and 9.786 = 6a0 + 14a1

By solving above equations we get, a0 = 0.0416, a1 = 0.6811

Q) Define Correlation.
Correlation is a term that is a measure of the strength of a linear
relationship between two quantitative variables (e.g., height, weight).

Positive correlation is a relationship between two variables in which both

variables move in the same direction and viceversa is Negative correlation.

eg. postivie correlation: the more you exercise, the more calories you will
burn.
 Q) Define Covariance.

The covariance of two variables x and y in a dataset measures how they are
linearly related. A positive covariance would indicate a positive linear
relationship between the variables and negative covariance indicates
opposite.

Eg. Find the covariance between x and y where x = (2.1, 2.2, 3.6, 4.0) and y
= (8, 10, 12, 14).

Q) Explain in brief about correlation coefficient.

The correlation coefficient of two variables in a dataset equals to their

covariance divided by product of their individual standard deviations.

The range of correlation coefficient is -1 to +1.

Eg. Find the correlation coefficient for above x & y.