Natural language processing basics in Intro to Computing - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What does tokenization do in NLP?

In natural language processing, what is the main purpose of tokenization?

A It converts text into numerical vectors directly.

B It translates text from one language to another.

C It removes all punctuation from the text.

D It splits text into smaller pieces like words or sentences.

💡 Hint

Think about how you break a sentence into parts before understanding it.

Tokenization breaks text into smaller units called tokens, usually words or sentences, which are easier to analyze.

🔍 Analysis

intermediate

Output of simple word count code

What is the output of this Python code that counts word frequencies?

Intro to Computing

from collections import Counter
text = 'apple banana apple orange banana apple'
word_counts = Counter(text.split())
print(word_counts)

A Counter({'apple': 3, 'banana': 2, 'orange': 1})

B Counter({'banana': 3, 'apple': 2, 'orange': 1})

C {'apple': 3, 'banana': 2, 'orange': 1}

D Counter({'apple': 1, 'banana': 1, 'orange': 1})

💡 Hint

Count how many times each word appears in the text.

The Counter counts each word's frequency: 'apple' appears 3 times, 'banana' 2 times, and 'orange' once.

❓ Model Choice

advanced

Best model for sentiment analysis

You want to build a model to classify movie reviews as positive or negative. Which model type is best suited for this task?

A Convolutional Neural Network (CNN) designed for image recognition

B K-Means clustering for grouping similar reviews

C Recurrent Neural Network (RNN) or Transformer model for text sequences

D Linear regression predicting review length

💡 Hint

Think about models that understand sequences of words.

RNNs and Transformers are designed to process sequences like text, making them suitable for sentiment analysis.

❓ Metrics

advanced

Choosing the right metric for imbalanced text classification

You trained a spam detection model where spam messages are rare. Which metric is best to evaluate your model's performance?

A Precision and recall, to balance false positives and false negatives

B Accuracy, because it shows overall correct predictions

C Mean squared error, to measure prediction errors

D Confusion matrix size, to count total classes

💡 Hint

Think about what matters when one class is much smaller than the other.

Precision and recall help measure how well the model detects rare spam messages without too many mistakes.

🔍 Analysis

expert

Why does this text preprocessing code raise an error?

Consider this Python code snippet for lowercasing and removing punctuation from text. Why does it raise an error?

Intro to Computing

import string
text = 'Hello, World!'
clean_text = text.lower().translate(str.maketrans('', '', string.punctuation))
print(clean_text)

A lower() cannot be chained with translate(), causing an AttributeError

B translate() expects a translation table, but string.punctuation is a string, causing a TypeError

C string.punctuation is empty, so translate() does nothing

D print() is missing parentheses, causing a SyntaxError

💡 Hint

Check what the translate() method needs as input.

translate() requires a translation table created by str.maketrans(), not a string like string.punctuation.