wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Page 1

Total questions: 150

Worksheet time: 1hrs 15mins

Name
Class
Date
1.

Machine Learning is a field that enables computers to learn without being

a)

Programmed

b)

Explicitly programmed

c)

Supervised

d)

Controlled

2.

Learning from labeled data is called

a)

Unsupervised learning

b)

Supervised learning

c)

Reinforcement learning

d)

Deep learning

3.

Candidate Elimination algorithm is used in

a)

Clustering

b)

Concept learning

c)

Regression

d)

Reinforcement

4.

Perceptron is a

a)

Linear classifier

b)

Clustering algorithm

c)

Regression model

d)

Optimization method

5.

Linear regression is used for

a)

Classification

b)

Clustering

c)

Prediction of continuous values

d)

Feature selection

6.

Version space represents

a)

All possible hypotheses

b)

Training data

c)

Testing data

d)

Output values

7.

Perceptron works only when data is

a)

Non-linear

b)

Linearly separable

c)

Random

d)

Unlabeled

8.

Gradient descent is used to minimize

a)

Accuracy

b)

Error

c)

Clusters

d)

Data size

9.

Concept learning is treated as a

a)

Search problem

b)

Clustering problem

c)

Regression problem

d)

Optimization problem

10.

Training data contains

a)

Only inputs

b)

Only outputs

c)

Input-output pairs

d)

Random values

11.

Linear discriminants are used for

a)

Clustering

b)

Classification

c)

Regression

d)

Optimization

12.

Which is supervised learning?

a)

K-means

b)

PCA

c)

Decision tree

d)

Apriori

13.

Perceptron learning updates

a)

Weights

b)

Inputs

c)

Outputs

d)

Labels

14.

Hypothesis space contains

a)

All possible models

b)

Training samples

c)

Errors

d)

Outputs

15.

Overfitting occurs when model performs well on

a)

Training data only

b)

Test data only

c)

New data

d)

All data

16.

Learning from labeled data is called (a)   learning.

17.

Perceptron is a (a)   classifier.

18.

Candidate Elimination algorithm maintains (a)   sets.

19.

Linear regression predicts (a)   values.

20.

Concept learning is viewed as a (a)   problem.

21.

Version space lies between (a)   and specific hypotheses.

22.

Perceptron works on (a)   separable data.

23.

Training data contains input and (a)   .

24.

Gradient descent minimizes (a)   .

25.

Supervised learning requires (a)   data.

26.

Hypothesis space is a set of (a)   .

27.

Classification predicts (a)   labels.

28.

Regression predicts (a)   outputs.

29.

Linear discriminants are used for (a)   .

30.

Perceptron is a single layer (a)   network.

31.

MLP stands for

a)

Multi Layer Perceptron

b)

Maximum Likelihood Predictor

c)

Multi Level Processor

d)

Modular Learning Protocol

32.

Backpropagation is used to

a)

initialize random weights

b)

update weights

c)

compute input features

d)

normalize outputs

33.

RBF stands for

a)

Rectified Bilinear Function

b)

Random Basis Filter

c)

Radial Basis Function

d)

Residual Backward Flow

34.

SVM is mainly used for

a)

clustering

b)

regression only

c)

classification

d)

dimensionality reduction

35.

Error in neural networks is propagated

a)

forward

b)

laterally

c)

backward

d)

randomly

36.

Kernel trick is used in

a)

PCA

b)

K-means

c)

SVM

d)

Naive Bayes

37.

Hidden layers are present in

a)

single-layer perceptron

b)

logistic regression

c)

MLP

d)

k-NN

38.

Curse of dimensionality mainly affects (a)   .

39.

Activation functions are used in (a)   .

40.

SVM maximizes (a)   .

41.

Sigmoid is an (a)   .

42.

Training in MLP is (a)   .

43.

RBF networks use (a)   .

44.

Hyperplane is a concept in (a)   .

45.

Backpropagation minimizes (a)   .

46.

Backpropagation updates (a)   .

47.

MLP contains input, hidden and (a)   layers.

48.

SVM finds optimal (a)   .

49.

RBF networks use (a)   functions.

50.

Kernel functions map data to (a)   dimensions.

51.

Learning in neural networks is (a)   .

52.

Error is propagated (a)   .

53.

SVM is a (a)   classifier.

54.

Hidden neurons use (a)   functions.

55.

Training data is required for (a)   learning.

56.

Overfitting can occur in (a)   networks.

57.

RBF network is a type of (a)   network.

58.

SVM can handle (a)   separable data.

59.

Backpropagation uses (a)   descent.

60.

Output of neuron depends on (a)   .

61.

Decision trees are used for

a)

clustering only

b)

classification only

c)

regression only

d)

classification and regression

62.

Entropy measures

a)

purity

b)

impurity

c)

margin

d)

variance

63.

Information gain is based on

a)

entropy

b)

Gini index

c)

variance reduction

d)

margin maximization

64.

CART stands for

a)

Classification and Regression Trees

b)

Cluster Analysis and Rule Trees

c)

Conditional Attribute Ranking Trees

d)

Combined and Averaged Regression Techniques

65.

Bagging is an

a)

optimization algorithm

b)

ensemble method

c)

feature selection technique

d)

distance metric

66.

Boosting combines

a)

weak learners

b)

strong learners

c)

random features

d)

bootstrap samples

67.

K-means is

a)

supervised learning

b)

unsupervised learning

c)

reinforcement learning

d)

semi-supervised learning

68.

Random Forest uses

a)

boosting

b)

stacking

c)

bagging

d)

pruning

69.

Gini index measures

a)

variance

b)

purity

c)

impurity

d)

entropy

70.

Pruning is used to reduce

a)

training time

b)

underfitting

c)

overfitting

d)

feature dimensionality

71.

KNN is based on

a)

probability

b)

information gain

c)

distance

d)

gradient

72.

Clustering uses

a)

labeled data

b)

unlabeled data

c)

balanced data

d)

oversampled data

73.

Leaf node represents

a)

root

b)

split attribute

c)

class label

d)

entropy

74.

Ensemble methods improve

a)

training data size

b)

accuracy

c)

feature count

d)

label count

75.

GMM is used for

a)

regression

b)

classification only

c)

clustering

d)

feature scaling

76.

Decision trees are used for (a)   .

77.

Entropy measures (a)   .

78.

Information gain selects the best (a)   .

79.

CART stands for (a)   .

80.

Combining models is called (a)   learning.

81.

Bagging reduces (a)   .

82.

Boosting reduces (a)   .

83.

K-means is an (a)   algorithm.

84.

Random Forest is based on (a)   .

85.

Pruning reduces (a)   .

86.

KNN uses (a)   metric.

87.

Leaf nodes give (a)   .

88.

Gini index measures (a)   .

89.

Clustering groups (a)   data.

90.

Gaussian Mixture Models are used for (a)   .

91.

PCA stands for

a)

Principal Component Analysis

b)

Partial Correlation Analysis

c)

Probabilistic Component Aggregation

d)

Principal Cluster Analysis

92.

PCA is used for

a)

hyperparameter tuning

b)

dimensionality reduction

c)

feature scaling

d)

model selection

93.

LDA stands for

a)

Latent Dirichlet Allocation

b)

Linear Discriminant Analysis

c)

Local Density Approximation

d)

Logistic Discriminant Analysis

94.

ICA stands for

a)

Independent Component Analysis

b)

Integrated Cluster Analysis

c)

Incremental Component Averaging

d)

Independent Correlation Assessment

95.

Isomap is a

a)

linear classifier

b)

non-linear reduction technique

c)

clustering algorithm

d)

feature selection method

96.

LLE stands for

a)

Locally Linear Embedding

b)

Least Likelihood Estimation

c)

Linear Laplacian Encoding

d)

Localized Logistic Embedding

97.

Genetic Algorithms belong to

a)

instance-based learning

b)

evolutionary learning

c)

reinforcement learning

d)

supervised learning

98.

Crossover is a

a)

gradient estimator

b)

regularization trick

c)

genetic operator

d)

clustering heuristic

99.

Complete the statement: Mutation introduces (a)   .

100.

Complete the statement: PCA works on (a)   .

101.

Complete the statement: LDA maximizes (a)   .

102.

Complete the statement: Dimensionality reduction removes (a)   .

103.

Complete the statement: Genetic algorithms are inspired by (a)   .

104.

Complete the statement: Optimization in ML uses (a)   .

105.

Complete the statement: Feature extraction is done by (a)   .

106.

Fill in the blank: PCA reduces (a)   .

107.

Fill in the blank: LDA is a (a)   method.

108.

Fill in the blank: ICA separates (a)   signals.

109.

Fill in the blank: LLE is a (a)   technique.

110.

Fill in the blank: Genetic algorithms use (a)   operators.

111.

Fill in the blank: Crossover and mutation are (a)   operations.

112.

Fill in the blank: PCA uses (a)   of data.

113.

Fill in the blank: Isomap preserves (a)   distance.

114.

Fill in the blank: Dimensionality reduction improves (a)   .

115.

Fill in the blank: Factor analysis is used for (a)   .

116.

Fill in the blank: Mutation introduces (a)   .

117.

Fill in the blank: Selection chooses (a)   individuals.

118.

Fill in the blank: Genetic algorithms are (a)   search techniques.

119.

Fill in the blank: PCA produces (a)   components.

120.

Fill in the blank: LDA maximizes (a)   separation.

121.

Choose the correct option: RL stands for ________.

a)

Reinforcement Learning

b)

Representation Learning

c)

Relational Logic

d)

Regression Learning

122.

Choose the correct option: RL learns by ________.

a)

trial and error

b)

memorization of labels

c)

deterministic planning

d)

unsupervised clustering

123.

Choose the correct option: MCMC stands for ________.

a)

Markov Chain Monte Carlo

b)

Multiple Constraint Monte Carlo

c)

Markov Controlled Model Class

d)

Matrix Chain Monte Computation

124.

Choose the correct option: HMM stands for ________.

a)

Hidden Markov Model

b)

Hierarchical Mixture Model

c)

Hybrid Markov Mechanism

d)

Hidden Memory Machine

125.

Choose the correct option: Bayesian networks are ________.

a)

probabilistic models

b)

deterministic rule sets

c)

distance metrics

d)

optimization solvers

126.

Choose the correct option: RL uses ________.

a)

reward signals

b)

class labels

c)

genetic crossover

d)

kernel tricks

127.

Choose the correct option: MCMC is used for ________.

a)

sampling

b)

feature scaling

c)

graph layout

d)

error correction

128.

Choose the correct option: Graphical models represent ________.

a)

probabilistic relations

b)

geometric transforms

c)

deterministic pipelines

d)

sorting procedures

129.

Choose the correct option: HMM is used for ________.

a)

sequence modeling

b)

image denoising

c)

nonlinear control

d)

spatial indexing

130.

Choose the correct option: In reinforcement learning, a policy defines ________.

a)

actions

b)

transitions between labels

c)

distance metrics

d)

network topology

131.

Choose the correct option: A Markov process depends on ________.

a)

current state

b)

entire history

c)

next reward only

d)

random seed

132.

Choose the correct option: Tracking methods commonly use ________.

a)

HMM

b)

k-d trees

c)

support vector regression

d)

Fourier descriptors

133.

Choose the correct option: Bayesian networks are ________.

a)

directed graphs

b)

undirected trees

c)

fully connected meshes

d)

acyclic lattices

134.

Choose the NB: correct option: An RL agent interacts with ________.

a)

environment

b)

labeling function

c)

static dataset only

d)

kernel

135.

Choose the correct option: Monte Carlo methods use ________.

a)

random sampling

b)

deterministic integration

c)

grid search only

d)

symbolic algebra

136.

RL stands for (a)   .

137.

RL learns through (a)   and error.

138.

MCMC is used for (a)   .

139.

HMM stands for (a)   .

140.

Bayesian networks are (a)   models.

141.

Reward guides (a)   learning.

142.

Graphical models represent (a)   .

143.

Markov process depends on (a)   state.

144.

Policy maps states to (a)   .

145.

HMM is used for (a)   data.

146.

Tracking methods use (a)   .

147.

Monte Carlo uses (a)   sampling.

148.

RL agent interacts with (a)   .

149.

Proposal distribution is used in (a)   .

150.

Markov Random Fields are (a)   models.