[4]

Q.6 Attempt any two: Total No. of Questions: 6 Total No. of Printed Pages:4
i. Explain the process of sentiment analysis and discuss its importance 5
in various real-world applications. Provide examples of how Enrollment No......................................
sentiment analysis can be used in business, social media, and
Faculty of Engineering
customer feedback analysis.
ii. Discuss the challenges involved in word sense disambiguation and 5 End Sem Examination May-2024
describe common techniques used to address these challenges. IT3EA06 Natural Language Processing
Provide examples to illustrate how word sense disambiguation can Programme: B.Tech. Branch/Specialisation: IT
improve the accuracy of natural language processing tasks such as Duration: 3 Hrs. Maximum Marks: 60
information retrieval and machine translation.
iii. Explain the concept of text classification and discuss its role in 5 Note: All questions are compulsory. Internal choices, if any, are indicated. Answers of
organizing and categorizing large volumes of textual data. Describe Q.1 (MCQs) should be written in full instead of only a, b, c or d. Assume suitable data if
common approaches to text classification, such as supervised necessary. Notations and symbols have their usual meaning.
learning algorithms and deep learning models. Q.1 i. What is the main approach of Natural Language Processing (NLP)? 1
(a) Statistical analysis (b) Symbolic rules
****** (c) Neural networks (d) All of these
ii. Which of the following best describes regular expression? 1
(a) A method for defining patterns in text data
(b) A programming language for AI development
(c) An algorithm for speech recognition
(d) A model for sentiment analysis
iii. Which technique is used to break down text into smaller meaningful 1
units such as words or subwords?
(a) Morphological parsing (b) Tokenization
(c) Feature extraction (d) Part of speech tagging
iv. Which approach to Part of Speech (POS) tagging relies on 1
predefined linguistic rules and patterns?
(a) Rule-Based POS Tagging
(b) Stochastic POS Tagging
(c) Transformation-Based Tagging
(d) None of these
v. Which method in speech processing utilizes Bayesian principles to 1
model and correct spelling errors in text?
(a) Spelling error
(b) Minimum edit distance
(c) Bayesian method to spelling
(d) Bayesian method of pronunciation variation

P.T.O.
 [2] [3]
vi. Which model in speech processing is commonly used to capture the 1 Q.3 i. Explain the concept of tokenization in text preprocessing. 2
uncertainty and variability in speech data? ii. Discuss the significance of Finite State Transducers (FSTs) in 8
(a) Phonological rules and transducer morphological analysis and parsing, highlighting their role,
(b) Probabilistic models advantages, and applications in natural language processing.
(c) Minimum edit distance OR iii. Compare and contrast three approaches to Part of Speech (POS) 8
(d) Bayesian method of pronunciation variation tagging: Rule-Based POS Tagging, Stochastic POS Tagging, and
vii. What is used to measure the uncertainty or unpredictability of a 1 Transformation-Based Tagging. Discuss the underlying principles,
language model? advantages, limitations, and typical applications of each approach in
(a) N-Grams (b) Perplexity natural language processing.
(c) Smoothing (d) Entropy
viii. Which technique is used to handle the problem of unseen N-Grams 1 Q.4 i. Explain the concept of minimum edit distance in the context of 3
in language models? speech processing and its application in tasks such as speech
(a) Simple N-Gram (b) Smoothing recognition or spelling correction.
(c) Parsing (d) Treebank ii. Discuss the application of Bayesian methods in modeling and 7
ix. Which application in natural language processing focuses on 1 addressing pronunciation variation in speech processing. Provide
determining the sentiment or emotional tone of a piece of text? examples of how Bayesian principles are used to handle uncertainty
(a) Sentiment analysis (b) Spelling correction and variability in spoken language.
(c) Machine translation (d) Text classification OR iii. Discuss the role of probabilistic models in speech processing, 7
x. Which task involves determining the intended meaning of a word 1 highlighting their significance in tasks such as speech recognition
based on its context in a sentence? and phonetic analysis. Provide examples of probabilistic models
(a) Sentiment analysis commonly used in these applications and explain how they address
(b) Word sense disambiguation the challenges of variability and uncertainty in speech data.
(c) Text classification
(d) Question answering system Q.5 i. Explain the concept of entropy in the context of NLP language 4
models and its relationship with the predictability of a text. Provide
Q.2 i. Define ambiguity in the context of Natural Language Processing 2 examples to illustrate how entropy is calculated and interpreted in
(NLP). N-Gram models.
ii. Compare and contrast the main approaches of Natural Language 3 ii. Compare and contrast the concepts of smoothing and backoff in 6
Processing (NLP), namely statistical analysis, symbolic rules, and N-Gram language models. Discuss their respective roles in
neural networks. addressing the problem of data sparsity and improving the
iii. Discuss the importance of formal languages and natural languages 5 performance of language models. Provide examples to illustrate
in the context of NLP, highlighting their differences and how they their usage in practical applications.
intersect. OR iii. Describe the significance of treebanks in statistical parsing and 6
OR iv. Explain the role of models and algorithms in the field of Natural 5 natural language processing. Discuss how treebanks are constructed,
Language Processing (NLP), highlighting their significance in annotated, and utilized in training and evaluating parsing
addressing language-related challenges. algorithms. Provide examples of commonly used treebanks and
explain their role in advancing research in syntactic analysis and
parsing.

P.T.O.
 [4]

Marking Scheme Application 2 Marks

OR iii. Compare and contrast three approaches to Part of Speech (POS) 8
NATURAL LANUAGE PROCESSING (NLP) IT3EA06 tagging: Rule-Based POS Tagging, Stochastic POS Tagging, and
Transformation-Based Tagging. Discuss the underlying principles,
Q.1 i) D) All of the above 1 advantages, limitations, and typical applications of each approach
ii) A) A method for defining patterns in text data 1 in natural language processing.
iii) B) Tokenization 1
iv) A) Rule-Based POS Tagging 1 Three approaches 3 Marks
v) C) Bayesian Method to Spelling 1 Principle advantage 2 Marks
vi) B) Probabilistic Models 1 Applications 3 Marks
vii) B) Perplexity 1
viii) B) Smoothing 1 Q.4 i. Explain the concept of Minimum Edit Distance in the context of 3
ix) A) Sentiment Analysis 1 speech processing and its application in tasks such as speech
x) B) Word Sense Disambiguation 1 recognition or spelling correction.

Q.2 i. Define ambiguity in the context of Natural Language Processing 2 Concept 1.5 Marks
(NLP). Application 1.5 Marks
ii. Compare and contrast the main approaches of Natural Language 3 ii. Discuss the application of Bayesian methods in modeling and 7
Processing (NLP), namely statistical analysis, symbolic rules, and addressing pronunciation variation in speech processing. Provide
neural networks. examples of how Bayesian principles are used to handle
uncertainty and variability in spoken language.
1 Mark each
iii. Discuss the importance of formal languages and natural languages 5 Application of Bayesian 3.5 Marks
in the context of NLP, highlighting their differences and how they Example 3.5 Marks
intersect. OR iii. Discuss the role of probabilistic models in speech processing, 7
highlighting their significance in tasks such as speech recognition
Importance 2.5 Marks and phonetic analysis. Provide examples of probabilistic models
Differences 2.5 Marks commonly used in these applications and explain how they
OR iv. Explain the role of models and algorithms in the field of Natural 5 address the challenges of variability and uncertainty in speech
Language Processing (NLP), highlighting their significance in data.
addressing language-related challenges.
Probabilistic models 3 Marks
Role of model and algorithm 2.5 Marks Significance 2 Marks
Addressing 2.5 Marks Applications 2 Marks
Q.3 i. Explain the concept of tokenization in text preprocessing. 2 Q.5 i. Explain the concept of entropy in the context of NLP language 4
models and its relationship with the predictability of a text.
2 Marks Provide examples to illustrate how entropy is calculated and
ii. Discuss the significance of Finite State Transducers (FSTs) in 8 interpreted in N-Gram models.
morphological analysis and parsing, highlighting their role,
advantages, and applications in natural language processing. Concept 1.5 Marks
Entropy calculation 2.5 Marks
FST 4 Marks ii. Compare and contrast the concepts of smoothing and backoff in 6
Advantage 2 Marks N-Gram language models. Discuss their respective roles in

P.T.O.
 [2] [3]
addressing the problem of data sparsity and improving the
performance of language models. Provide examples to illustrate
their usage in practical applications.

Comparison 2 Marks
Roles in addressing data sparsity 3 Marks.
Example 1 Marks
OR iii. Describe the significance of Treebanks in statistical parsing and 6
natural language processing. Discuss how Treebanks are
constructed, annotated, and utilized in training and evaluating
parsing algorithms. Provide examples of commonly used
Treebanks and explain their role in advancing research in
syntactic analysis and parsing.

Significance of treebanks 2 Marks

Treebank construction 3 Marks
Example 1 Marks

Q.6 Attempt any two:

i. Explain the process of sentiment analysis and discuss its 5
importance in various real-world applications. Provide examples
of how sentiment analysis can be used in business, social media,
and customer feedback analysis.

Process 2.5 Marks

Example 2.5 Marks
ii. Discuss the challenges involved in word sense disambiguation and 5
describe common techniques used to address these challenges.
Provide examples to illustrate how word sense disambiguation can
improve the accuracy of natural language processing tasks such as
information retrieval and machine translation.

Word sense disambiguation and techniques 3 Marks

Example 2 Marks
iii. Explain the concept of text classification and discuss its role in 5
organizing and categorizing large volumes of textual data.
Describe common approaches to text classification, such as
supervised learning algorithms and deep learning models.

Text classification 1.5 Marks

Its role 1 Marks
Approaches 2.5 Marks

******

P.T.O.