Data extraction from text in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main purpose of Named Entity Recognition (NER) in data extraction?

Named Entity Recognition (NER) is a common technique in data extraction from text. What does NER primarily do?

A It identifies and classifies key information like names, dates, and locations in text.

B It translates text from one language to another.

C It summarizes long documents into short paragraphs.

D It generates new text based on input prompts.

💡 Hint

Think about what kind of information you want to pull out from a text document.

NER finds specific pieces of information such as names, dates, and places, which helps in extracting structured data from unstructured text.

❓ Predict Output

intermediate

Output of simple regex extraction code

What is the output of the following Python code that extracts all email addresses from a text?

Prompt Engineering / GenAI

import re
text = 'Contact us at [e-mail] or [e-mail].'
emails = re.findall(r'\b[\w.-]+@[\w.-]+\.\w+\b', text)
print(emails)

A ['[e-mail]', '[e-mail]']

B ['[e-mail] [e-mail]']

C ['support@example', 'sales@example']

D []

💡 Hint

Look at the regex pattern and what re.findall returns.

The regex matches email patterns and re.findall returns a list of all matches found in the text.

❓ Model Choice

advanced

Best model type for extracting structured data from unstructured text

You want to extract structured information like product names, prices, and dates from customer reviews. Which model type is best suited for this task?

A K-Means clustering for grouping similar texts

B Generative Adversarial Network (GAN) for data generation

C Convolutional Neural Network (CNN) for image classification

D Recurrent Neural Network (RNN) or Transformer-based model for sequence labeling

💡 Hint

Think about models that handle sequences and labeling tasks.

RNNs and Transformer models are designed to process sequences and can label each token, making them suitable for extracting structured data from text.

❓ Metrics

advanced

Evaluating extraction accuracy with precision and recall

You built a model to extract dates from text. On a test set, it found 80 dates, of which 60 were correct. The test set actually contains 100 dates. What are the precision and recall?

A Precision = 0.80, Recall = 0.60

B Precision = 0.60, Recall = 0.75

C Precision = 0.75, Recall = 0.60

D Precision = 0.60, Recall = 0.80

💡 Hint

Precision = correct found / total found; Recall = correct found / total actual.

Precision = 60/80 = 0.75; Recall = 60/100 = 0.60.

🔧 Debug

expert

Why does this extraction code raise an error?

Consider this Python code snippet for extracting phone numbers. Why does it raise an error?

Prompt Engineering / GenAI

import re
text = 'Call me at 123-456-7890 or 987-654-3210.'
pattern = r'\d{3}-\d{3}-\d{4}'
matches = re.match(pattern, text)
print(matches.group())

A The regex pattern is invalid and causes a SyntaxError.

B re.match only checks the start of the string, so it returns None causing an AttributeError.

C The text variable is empty, so no matches are found.

D matches.group() is called before re.match is imported.

💡 Hint

Check how re.match works compared to re.findall or re.search.

re.match tries to match the pattern only at the start of the string. Since the phone number is not at the start, it returns None, so calling group() causes an error.