NLP study Material

Module 1

 Identify two real-world scenarios where Named Entity Recognition is utilized.

 Describe different categories of entities recognized by NER with relevant
illustrations.
 Define the concept of natural language processing in your own words.
 What are stop words and why are they often excluded from processing?
 Cite two domains where NLP techniques are commonly applied.
 Clarify how precision differs from recall in the evaluation of search results.
 Briefly describe the purpose of the NLTK library.
 Explain what is meant by identifying multiple words as a single unit in tokenization.
 What is stemming and how is it useful in NLP?
 Explain what affixes are and their function in word formation.
 What is a lexicon in NLP, and why is it important?
 Why do we sometimes prefer tokenizing phrases instead of individual words?
 Describe the task of breaking text into separate sentences.
 What is the importance of segmenting text into sentences?
 Define morphology as it pertains to NLP tasks.
 Name the key types of morphology relevant in computational linguistics.
 How does natural language understanding differ from NLP more generally?
 Provide examples of standard corpora used in NLP research.
 Distinguish between dividing text into words and segmenting into sentences.
 Outline the steps taken to solve a typical NLP problem.
 Show how a sentence is broken into tokens, using a specific sentence.
 Describe the inner workings of a named entity recognition system.
 Explain the value of removing stop words and when it might be counterproductive.
 Illustrate how regular expressions work through a simple case.
 Discuss how dependency parsing helps analyze sentence structure.
 Create a regex for all strings using only ‘a’ and ‘b’ with even length.
 Build a regex for strings of length four starting with the letter 'a'.
 Write a regex that ensures a string includes at least one ‘a’.
 Compare the features and capabilities of NLTK and Spacy.
 Define and exemplify the Bag of Words model.
 How do regular expressions differ from regular grammar?
 Walk through both word and sentence tokenization with a hands-on example.
 What are the frequent issues faced in morphological analysis, and how can they be
solved?
 Explain and demonstrate the Minimum Edit Distance algorithm with “MAM” and
“MADAM”.
 Describe how NER and IE are used to address real-life tasks.
 Pick a real-world NLP problem and explain how it was solved step-by-step.
 Explain how to build a corpus tailored to a specific NLP problem.
 What does Information Extraction aim to achieve in NLP?
 List uses of sentiment and opinion mining with corresponding variations.
 List practical examples of where Information Retrieval techniques are applied.
 Define text normalization and why it matters in NLP preprocessing.
 How does normalization differ from tokenization? Provide justification with
examples.
 Why is POS tagging important? Explain with a real example.
 Point out limitations of a basic Top-Down parsing technique.
 Clarify the contrast between classification and prediction using an example.
 How are word and phrase tokenization techniques used differently?
 Name two real-life examples that show NLP’s role across industries.
 Identify all strings of length 5 or fewer from given regular expressions.
 Write regex for:
 All alphabet-only strings
 Lowercase strings ending with ‘b’
 Strings where every 'a' is surrounded by 'b'
 Describe how rule-based POS tagging works.
 Explain how a regular expression contrasts with regular grammar.
 Briefly explain what NLTK is and its purpose.
 Clarify what’s meant by Multi Word Tokenization.
 Describe how sentence segmentation operates in NLP.
 What does morphology focus on in computational linguistics?
 Share well-known corpus datasets used in NLP.
 What does word tokenization involve?
 Calculate the minimum number of operations to turn “ELEPHANT” into
“RELEVANT”.
 Determine the edit distance between "SUNDAY" and "SATURDAY".
 Identify the different branches of morphology in linguistics.
 Explain what stemming is and where it fits in the NLP pipeline.
 Define a corpus and explain its role in NLP.
 Differentiate stemming and lemmatization with examples.
 List and describe the major stages in the NLP pipeline.
 What is the role of a chatbot in NLP-driven applications?
 Summarize the main steps in text preprocessing and explain how to manage outliers.
 Provide an example highlighting issues with sentence segmentation.
 What are some fundamental tasks NLP systems typically perform?
 Describe what is meant by text cleanup and extraction, with examples.
 What is word sense ambiguity and how can it impact NLP?
 Write a short note on how the Bag of Words approach works.
 What is homonymy? Provide a brief explanation with an example.
 Define WordNet and its utility.
 Given a document and corpus stats, compute TF and IDF values.
 Relate the use of SVD, matrix completion, and matrix factorization in NLP.
 Present two use cases showing how regex supports NLP tasks.
 Why is it beneficial to use multiword tokens over single tokens? Provide illustrations.
 Differentiate natural language from formal language.
 Define lexicon and lexeme, and explain the relationships between different lexemes.
 What makes bottom-up parsers more efficient than top-down ones?
  (Question skipped in original numbering; continue to next)
 Describe the Skip-gram model and its approach to learning embeddings.
 Clarify how TF-IDF scoring works in search ranking.
 Tokenize and tag a sample sentence.
 Identify different pronunciations and parts-of-speech of a given word.
 Use an edit distance grid to transform “intention” to “execution”.
 Why is creating corpora essential in NLP research?
 Discuss the role of regex in managing and exploring text data.
 Explain how WordNet supports understanding word relationships in NLP.
 Define pragmatic ambiguity and explain how it affects interpretation.
 Describe what kind of strings are matched by these regex patterns:
a. [a-zA-Z]+
b. [A-Z][a-z]*
 Extract email addresses from a sample text using regex.
 Analyze a faulty regex intended to capture uppercase letters and digits. Suggest a fix.
 Build a regex capable of detecting three different date formats in text.
 Calculate edit distance between “MAMA” and “MADAAM”.
 Find the transformation steps for turning “kitten” into “sitting” with edit cost 1

Module 2

 How would you define a language model in NLP?

 Explain how an n-gram model works, using a simple example.
 Mention two key differences between bigram and trigram models.
 What is the chain rule of probability, and how is it used in language modeling?
 When applying the bigram model, what simplifications do we introduce into
probability computation?
 Define the Markov assumption and explain its role in NLP.
 What purpose does maximum likelihood estimation serve in statistical NLP?
 Given a word and its preceding word, how would you calculate normalized bigram
probabilities?
 What does relative frequency mean in the context of n-gram models?
 Identify the core elements that make up a semantic interpretation system.
 What is lexical ambiguity and how does it affect language understanding?
 Provide an explanation and example of semantic ambiguity.
 What is syntactic ambiguity, and how does it arise in sentences?
 Why is it important to represent meaning in computational language processing?
 Explain the major contrast between lexical and semantic analysis.
 List two tools commonly used for creating or analyzing language models.
 Discuss various part-of-speech attributes with suitable examples.
 What is extrinsic evaluation in language models, and what challenges are associated
with it?
 Demonstrate how path-based similarity is used to compare word meanings.
 Define and differentiate homonymy, polysemy, and synonymy with examples.
 How can WordNet be used to derive meaning from text corpora?
 Explain how computational semantics is implemented in NLP systems.
 What are the drawbacks of basic path similarity, and how does information content
improve it?
 How does the extended Lesk algorithm work? Include a use case.
 Highlight the differences between rule-based and stochastic part-of-speech tagging
techniques.
 What are the features and advantages of stochastic POS tagging?
 Describe rule-based POS tagging and list its key properties.
 Give examples of how n-gram models assist in predicting the next word in a
sentence.
 What is transformation-based tagging and how does it function?
 Compare structured and unstructured data using NLP examples.
 What is semi-structured data, and how does it differ from structured data? Provide
an example.
 How can supervised learning methods help in categorizing text?
 List practical applications of emotion detection in real-world scenarios.
 Suppose you work for a food delivery company and need to survey competitors—
how would you carry this out using NLP techniques?
 Describe a traditional search model and illustrate it with a diagram.
 Why is tagging parts of speech vital for natural language tasks?
 What is meant by “vocabulary” in NLP and why does it matter?
 Define Information Extraction and state its importance.
 What is morphological parsing, and what are its major stages?
 Clarify the idea of a Bag-of-Words and give an example of its usage.
 How do formal and natural languages differ from each other?
 Suppose you want to regenerate a document based on its word content. How would
you do this and what does it tell you about the topic?
 What does text parsing mean in an NLP context?
 How is sentiment analysis applied in analyzing market trends?
 Briefly describe the Hidden Markov Model and its role in sequence labeling.
 Explain a key advantage of using Latent Dirichlet Allocation (LDA) over Probabilistic
Latent Semantic Analysis (PLSA) for recommendations.
 How can building a recommender system using matrix factorization be considered a
regression task?
 Name two common POS tagging methods used in computational linguistics.
 What is WordNet and what role does it play in NLP?
 Explain the hierarchy of word relationships as represented in WordNet.
 How are morphological operations implemented in NLP applications?
 Define hypernyms, hyponyms, and heteronyms, and explain with examples.
 Compare the advantages and disadvantages of Skip-gram and CBOW models.
 Describe how text classification works using the Naïve Bayes algorithm.
 How can a Naïve Bayes model be applied to collaborative filtering tasks?
 Clarify the difference between lexical and semantic analysis in NLP.
 Define what n-grams are and explain their use in language processing.
  What are word embeddings and how are they useful in NLP systems?
 Explain “vector semantics” and its role in interpreting word meaning.
 State a major limitation of the TF-IDF model in NLP applications.
 Show how regular expressions help in processing and cleaning text data.
 Describe the role of n-grams in language modeling, and illustrate how they detect
text patterns.
 Discuss how n-grams are applied in building classification systems.
 What problem does the unigram model present in extracting useful information?
 Define homographs and illustrate with a real-world example.
 How does the Levenshtein distance measure similarity between two words?
 What are heteronyms? Give one example.
 Define polysemy and illustrate it with a word that has multiple meanings.
 What are synonyms and antonyms? Provide examples of each.
 Given a small corpus and a bigram model, calculate the smoothed probability of a
word following another.
 Analyze the validity of the following:
o Are rule-based taggers deterministic?
o Do stochastic taggers depend on specific languages?

Module 3

 How can concepts like TF-IDF, data splitting (train/validation/test), and stop words
influence the performance of NLP-based machine learning models? What challenges
might arise, and how can they be addressed?
 What is meant by text classification in NLP?
 How can structured information be extracted from unstructured text?
 What are ad-hoc retrieval tasks in information systems?
 Which aspects of ad-hoc retrieval are typically addressed in IR research?
 What components are included in a basic Information Retrieval system?
 Define the purpose and structure of an inverted index.
 How can manually designed rules assist in categorizing text?
 List and explain a few machine learning techniques used for classifying text
documents.
 What are some known limitations of the Naive Bayes classifier in text tasks?
 Describe the outcome of applying the independence assumption in the multinomial
Naive Bayes model.
 Mention two scenarios where the bag-of-words approach can be effectively applied.
 What issue does maximum likelihood estimation present in multinomial Naive Bayes,
and how can this be mitigated?
 Using the concept of a confusion matrix, explain the working of a spam classifier.
 How is k-fold cross-validation used to test the robustness of a classifier?
 Identify real-world challenges faced by text classifiers and how to resolve them.
 What are the major categories of techniques used for text classification?
 List three performance metrics for evaluating classifiers. Explain each with suitable
examples.
 What evaluation parameters are used to assess the performance of a confusion matrix?
 Illustrate Word2Vec-based embeddings using a simple diagram.
 Explain how Doc2Vec differs from Word2Vec, with a labeled diagram.
 With examples, explain how vector semantics and probabilistic models help represent
sequences of words.
 Define the process of opinion mining in NLP.
 What are the major considerations when collecting feedback data for sentiment
analysis?
 What is intent analysis and where is it applied?
 Define emotion analysis and describe how it functions.
 Explain the mechanism behind emotion analytics and its impact on business or user
behavior.
 Why is the Naive Bayes model still powerful despite its name?
 Describe how the multinomial Naive Bayes algorithm works using a step-by-step
approach.
 Differentiate micro and macro averaging in classification metrics using an example.
 Explain three different techniques for opinion mining.
 What challenge arises in keyword-based IR when dealing with large documents?
 What are the key initial steps in the text preprocessing pipeline?
 What is the primary aim of an information retrieval system?
 In what different ways can a Bag-of-Words representation be leveraged for
classifying documents?
 Compare sentiment, emotion, and intent analysis in terms of purpose and output.
 How do companies utilize sentiment analysis to track product reception in the market?
 Name some practical implementations of emotion detection through recognition
techniques.
 Show how Naive Bayes can be applied step-by-step in a text classification scenario.
 List and explain the four essential stages involved in text normalization.
 Highlight specific real-world applications of automatic text classification.
 Define NER and explain its role in extracting meaningful elements from text.
 How is Named Entity Recognition useful in building NLP-based applications?
 Explain how k-fold validation helps in performance testing of classifiers.
 Discuss the core principles of NLP and why they are valuable in modern applications.
 What is ambiguity in language? Describe the different types found in NLP.
 List some key benefits of using automated text classification, with an example.
 What foundational concepts make up a semantic system in NLP?
 How does NLTK differ from Spacy in terms of features and use cases?
 Provide a detailed explanation of dependency parsing in syntactic analysis.
 What preprocessing tasks are typically performed before applying NLP techniques?
 Describe different industries where chatbots are being actively used.
 Using Levenshtein distance (insertion = 1, deletion = 1, substitution = 2), transform
“DOG” into “COW”.
 What is word embedding and how does it benefit various NLP applications?
 Clarify the difference between prediction and classification with practical examples.
 How does lexical ambiguity affect tasks such as translation or sentiment detection?
How does WordNet help?
 What is topic modeling in text analytics and what purpose does it serve?
  Using the given dataset, apply Naive Bayes to determine if an email with “Offer =
Yes”, “Win = Yes”, and “Money = Yes” is spam.
 Given a dataset of customer feedback, determine if a new feedback entry is “Positive”
using Naive Bayes.
 With a given weather dataset, predict whether someone will play tennis using Naive
Bayes based on the conditions:
o Outlook: Rain
o Temperature: Mild
o Humidity: High
o Wind: Strong

Module 4

 Break down the primary types of recommendation systems and how they are
categorized.
 How does a content-based recommendation engine function?
 Describe the operational principles behind collaborative filtering.
 List and briefly explain the key metrics used to evaluate recommendation systems.
 What is meant by a hybrid recommendation approach?
 Define conversational agents and explain their purpose.
 What is the role of summarization in processing large text data?
 How does item-based collaborative filtering differ from user-based methods?
 Provide some real-world use cases of topic modeling.
 Why is there a need to automatically summarize textual content?
 Justify the classification of a chatbot as a conversational system.
 What benefits does artificial intelligence bring to chatbot development?
 Explain the idea behind retrieval-based conversation models.
 Define a question-answering system and its purpose.
 Mention examples of commonly used question answering systems.
 How does the user-based collaborative filtering technique operate?
 Explain the approach of using IR for building a question answering system.
 Contrast the goals of information retrieval versus traditional web search.
 Use an example to describe how user ratings help shape a recommendation.
 What is sentiment analysis, and how can it be illustrated with an example?
 Describe the various types of recommendation engines available.
 Provide real-world scenarios that exemplify recommendation system usage.
 Explain two different types of conversational agents and their characteristics.
 Show with an example how collaborative filtering can generate recommendations.
 List the most frequent use cases where sentiment analysis is applied.
 What are the major steps involved in implementing Latent Dirichlet Allocation
(LDA)?
 Describe how sentiment from Twitter posts can be analyzed using NLP.
 What components define chatbot architecture in modern systems?
 Explain how summarizing multiple documents works.
 Define topic modeling and its significance in text analysis.
 What is extractive summarization and how is it implemented?
 Contrast extractive and abstractive summarization techniques.
 Categorize recommendation strategies and provide an example of each.
 Highlight various techniques for summarizing text content.
 What are the most relevant use-cases where recommendation engines are deployed?
 Differentiate content-based filtering from collaborative filtering.
 Outline the major steps of sentiment analysis in NLP.
 What makes LDA different from other topic modeling techniques?
 With a simple case, explain how abstractive summarization is done.
 (Originally incomplete in the source; continue with assumed context) Given a
collection of sentences, demonstrate how summarization techniques would condense
them.
 What are the key pros and cons of both collaborative and content-based filtering?
 How do TF-IDF, training-validation-test split, and stop word removal impact model
building in NLP? What are some common challenges, and how can they be
addressed?
 How do extractive and abstractive summarizers differ? How would you construct an
extractive summarization tool?
 Describe how models like GPT use pretraining to enhance performance on NLP tasks.
 How can a Naive Bayes model be adapted for use in collaborative filtering?
 Clarify how lexical and semantic analysis differ in language processing.
 How does NLP support sentiment analysis tasks?
 Provide a brief explanation of n-grams in the NLP context.
 Describe how n-grams are applied in practical NLP tasks.
 Define the term "data augmentation" in the context of NLP.
 What steps would you follow to build a recommendation system that works with text
data?
 Explain how CBOW and Skip-Gram are implemented as part of Word2Vec.
 What distinguishes collaborative filtering from content-based recommendation?
 Name a few real-world scenarios where recommendation systems are applied.
 Compare rule-based, retrieval-based, and generative chatbot architectures, evaluating
them by scalability, quality, and flexibility.
 How does ChatGPT leverage large-scale pretraining and transformer architecture to
deliver relevant replies?
 Provide an example to explain collaborative recommendation systems in action.
 List five diverse areas where NLP has real-world applications like education,
healthcare, or finance.
 Why is Natural Language Understanding (NLU) essential in chatbot design?
 Outline the architectural framework behind ChatGPT as an NLP model.
 Mention different ways to apply data augmentation in NLP-related tasks.
 Distinguish between information retrieval and a basic web search engine.
 Elaborate on the principles of item-based collaborative filtering.
 What is dialogue management and how is it relevant in chatbot development?
 How do large pre-trained models like GPT-3 enhance conversational systems?
 How does user-based collaborative filtering operate in recommender engines?
 Can statistical models support machine translation? If yes, give a short explanation.
 With a diagram, explain how both single and multi-document summarization
techniques function.
 Provide an example that demonstrates how abstractive summarization works.
 Highlight the strengths and weaknesses of collaborative vs. content-based filtering
approaches.