TA-1 DL-2 12-08-
2024
kashyapsp@[Link] Switch account
Your email will be recorded when you submit
this form
TA-1 DL-2 12-08-2024
30 Questions : 30 Min
Which of the following is not true 1 point
about Markov Networks?
They are undirected
They can represent conditional
independencies
They require directionality between
nodes
They can be used to model
relationships in probabilistic systems
In the context of Markov Networks, 1 point
what does the term "clique" refer
to?
A subset of nodes where all nodes are
directly connected
A sequence of independent nodes
A single node
A node with no edges
What is a "Factor" in a Markov 1 point
Network?
A conditional probability
A function that captures the aAnity
between variables
An isolated node
A measurement of independence
Which concept cannot be directly 1 point
represented in a Bayesian Network
but can in a Markov Network?
Direct relationships
Symmetric dependencies
Conditional independencies
Causal relationships
What is a Perfect Map in the 1 point
context of graphical models?
A graph where nodes are perfectly
independent
A graph that exactly captures all
independencies of a distribution
A complete graph
A graph with no edges
What is the purpose of the 1 point
partition function Z in a Markov
Network?
To normalize the factors
To calculate probabilities
To measure the network size
To determine independence
Which of the following is an 1 point
example of a Markov Network?
Naive Bayes
Hidden Markov Model
Ising Model
Decision Tree
In a Markov Network, which of the 1 point
following properties is used to
determine local independencies?
Directed edges
Markov Blanket
Conditional probability
Latent variables
What is the key difference between 1 point
Bayesian Networks and Markov
Networks?
Bayesian Networks are undirected;
Markov Networks are directed
Bayesian Networks require fewer
parameters
Markov Networks do not have a
direction of inNuence between nodes
Bayesian Networks cannot model
conditional independence
Which of the following is true for a 1 point
clique in a Markov Network?
It is always a set of two nodes
It cannot be factorized
It is a maximal complete subgraph
It represents independent nodes
What type of factor is used in 1 point
Markov Networks to represent
interactions?
Univariate
Joint Probability
Conditional Probability
Potential Function
What is the purpose of the Gibbs 1 point
distribution in Markov Networks?
To estimate network size
To model causal relationships
To represent the joint distribution as a
product of factors
To eliminate conditional dependencies
In which of the following scenarios 1 point
would you prefer a Markov Network
over a Bayesian Network?
When relationships are directional
When modeling causal relationships
When dealing with symmetric
dependencies
When modeling temporal sequences
How are the weights in a Markov 1 point
Network typically determined?
Through causal inference
Through learning from data
Through random assignment
Through expert knowledge only
Which algorithm is commonly used 1 point
for inference in Markov Networks?
Naive Bayes
Belief Propagation
Gradient Descent
Expectation Maximization
Which type of problem is Markov 1 point
Networks particularly well-suited
for?
Time series forecasting
ClassiWcation with independent
features
Modeling symmetric relationships
Predicting future events
What is the role of the Markov 1 point
Blanket in a Markov Network?
It determines the network's topology
It isolates a node from the rest of the
network
It represents all the variables
independent of a given node
It includes the neighbors of a node,
making it independent of the rest
Which of the following is a 1 point
potential advantage of Markov
Networks over Bayesian
Networks?
Easier to understand
Better at handling large datasets
More Nexible in representing symmetric
relationships
Requires fewer parameters
What does the term "graphical 1 point
model" generally refer to?
A model based on graphs like trees or
networks
A model that represents data in visual
format
A statistical model that uses graphs to
represent dependencies
A model that visualizes mathematical
functions
In Bayesian Networks, what does a 1 point
directed edge signify?
Symmetric dependency
Temporal sequence
Causal inNuence
Mutual exclusivity
What is the importance of the 1 point
"parent" nodes in a Bayesian
Network?
They are always independent
They conditionally inNuence their child
nodes
They do not affect their child nodes
They are isolated nodes
In a Bayesian Network, what does 1 point
the term "non-descendants" refer
to?
Nodes that have no effect on the node
in question
Nodes that are not parents or children
Nodes that are causally related
Nodes that are isolated
What is a potential drawback of 1 point
using Bayesian Networks?
They are computationally expensive to
construct
They cannot model symmetric
relationships
They require a lot of data to be accurate
They are not interpretable
What is a "joint probability 1 point
distribution" in the context of
graphical models?
The probability distribution over all
variables in the network
The probability of two variables being
independent
The probability of a single variable
The distribution of factors across
cliques
What is the relationship between 1 point
cliques and factors in a Markov
Network?
Each factor corresponds to a clique in
the network
Factors are independent of cliques
Cliques determine the edges in a
Markov Network
Factors are only relevant in Bayesian
Networks
A company wants to model the 1 point
interactions between different
departments (Sales, Marketing, HR,
etc.) to understand how an issue in
one department might affect others
without assuming any directional
influence. Which model should they
use?
Bayesian Network
Decision Tree
Markov Network
Linear Regression
In a medical diagnosis system, you 1 point
need to predict whether a patient
has a certain disease based on
symptoms, where the presence of
one symptom increases the
likelihood of another. Which model
would be best to represent this
scenario?
Bayesian Network
Logistic Regression
Markov Network
Random Forest
You are designing a diagnostic 1 point
system for a car's engine, where a
fault in one part may directly cause
faults in other connected parts.
Which type of network should you
use to model this causal
relationship?
Bayesian Network
Markov Network
Random Forest
Clustering
A recommendation engine needs 1 point
to suggest products to users based
on the combination of products they
have already purchased, with the
assumption that certain products
are more likely to be purchased
together. Which model best
captures these relationships?
Markov Network
Bayesian Network
Support Vector Machine
Linear Regression
In a gene regulatory network, 1 point
genes influence each other’s
expression, but not necessarily in a
directed manner. What model would
be best to capture the
interdependencies among these
genes?
Bayesian Network
Markov Network
Neural Network
Linear Regression
Back Submit Clear form
Never submit passwords through Google Forms.
This form was created inside of RKNEC. - Contact form
owner
Forms