Lemmatization in NLP - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Lemmatization with POS Tag

What is the output of the following Python code using NLTK's WordNetLemmatizer when lemmatizing the word 'better' as an adjective?

NLP

from nltk.stem import WordNetLemmatizer
lemmatizer = WordNetLemmatizer()
result = lemmatizer.lemmatize('better', pos='a')
print(result)

A "better"

B "bet"

C "good"

D "bettered"

💡 Hint

Consider how lemmatization uses part of speech to find the base form.

When lemmatizing 'better' as an adjective (pos='a'), the lemmatizer returns its base form 'good'. Without specifying pos, it would return 'better'.

🧠 Conceptual

intermediate

Purpose of Lemmatization in Text Processing

Which of the following best describes the main purpose of lemmatization in natural language processing?

A To translate text from one language to another

B To remove all punctuation and special characters from text

C To split text into individual words or tokens

D To reduce words to their base or dictionary form considering context

💡 Hint

Think about how lemmatization differs from simple word cutting.

Lemmatization reduces words to their base or dictionary form (lemma) by considering the word's meaning and part of speech, unlike stemming which may just cut off endings.

❓ Metrics

advanced

Evaluating Lemmatization Impact on Text Classification

You apply lemmatization to your training and test datasets before training a text classifier. Which metric is best to compare model performance before and after lemmatization?

A Accuracy on the test set

B Number of unique tokens in the vocabulary

C Training loss only

D Time taken to train the model

💡 Hint

Focus on how well the model predicts unseen data.

Accuracy on the test set measures how well the model generalizes to new data, showing if lemmatization improved understanding.

🔧 Debug

advanced

Identifying the Error in Lemmatization Code

What error will this code raise when run? from nltk.stem import WordNetLemmatizer lemmatizer = WordNetLemmatizer() word = 'running' result = lemmatizer.lemmatize(word) print(result)

NLP

from nltk.stem import WordNetLemmatizer
lemmatizer = WordNetLemmatizer()
word = 'running'
result = lemmatizer.lemmatize(word)
print(result)

A Prints 'run' without error

B Prints 'running' without error

C Raises TypeError due to missing pos argument

D Raises NameError because WordNetLemmatizer is not imported

💡 Hint

Check the default behavior of lemmatize without pos argument.

Without specifying pos, lemmatize treats the word as a noun, so 'running' stays 'running'. No error occurs.

❓ Model Choice

expert

Choosing the Best Lemmatization Approach for Ambiguous Words

You want to lemmatize words in a sentence where words can have multiple parts of speech (e.g., 'saw' as noun or verb). Which approach will give the most accurate lemmatization?

A First perform POS tagging, then lemmatize each word with its POS tag

B Use a stemmer instead of a lemmatizer

C Lemmatize all words assuming they are nouns

D Lemmatize words without specifying POS tags

💡 Hint

Think about how context helps decide the correct base form.

POS tagging before lemmatization helps identify the correct part of speech, so the lemmatizer can find the right base form for ambiguous words.