Project 3 Midterm Review

Why is Natural Language Processing (NLP) especially difficult for computers?

• A) Computers have infinite memory

• B) Human language is ambiguous and context-dependent

• C) Programming languages are too complex

• D) Data in NLP is usually numeric

What is the main purpose of tokenization in NLP?

• Encrypt text for security

• Convert text into lower case

• Split text into smaller units like words

• Translate text into another language

This is a binary classification model. What number should you put in the blank?
output = Dense(_, activation='sigmoid')(x)

(a)  

Which of the following is considered a "stop word"?

• beautiful

• happiness

• the

• information

Why are stop words sometimes removed in NLP tasks?

• They are always grammatically incorrect

• They contain confidential information

• They don’t carry significant meaning

• They are used only in spoken language

What is stemming in NLP?

• Removing grammatical errors from text

• Splitting text into individual characters

• Reducing a word to its base form (physical letter base)

• Counting the number of syllables in a word

Imagine the word “better” - if you reduce the word to “good”, what have you done?

• Lemmatization

• Tokenization

• NER

• Stemming

What might be a downside of stemming?

• It returns full dictionary words

• It can be too accurate

• Words can lose their meaning

• It uses a lot of memory

What is a function of the Bag of Words (BoW) model in NLP?

• Predicting the next word in a sentence

• Representing text as numerical feature vectors

• Translating text from one language to another

• Detecting sarcasm in text

How does the BoW model represent text data?

• Using word embeddings and context

• By counting the frequency of words in a document

• By converting each word to its lemma

• By analyzing grammar rules