LDA with Gensim in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a dictionary from tokenized documents.

NLP

from gensim.corpora import Dictionary

docs = [['apple', 'banana', 'apple'], ['banana', 'orange']]
dictionary = Dictionary([1])

Drag options to blanks, or click blank then click option'

A['apple', 'orange']

B['apple', 'banana']

C['banana', 'orange']

Ddocs

Attempts:

3 left

💡 Hint

Common Mistakes

Passing a single list of words instead of a list of tokenized documents.

Passing a string instead of a list.

The Dictionary class expects a list of tokenized documents to create the mapping of word IDs.

2fill in blank

medium

Complete the code to convert documents into a bag-of-words corpus using the dictionary.

NLP

corpus = [[1] for doc in docs]

Drag options to blanks, or click blank then click option'

Adictionary.doc2bow(docs)

Bdictionary.doc2bow('doc')

Cdictionary.doc2bow(doc)

Ddictionary.doc2bow(['doc'])

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the whole list of documents instead of a single document.

Passing a string instead of a list of tokens.

The doc2bow method converts a tokenized document into a bag-of-words format using the dictionary.

3fill in blank

hard

Fix the error in the code to create an LDA model with 2 topics.

NLP

from gensim.models import LdaModel

lda = LdaModel(corpus=corpus, id2word=[1], num_topics=2, random_state=42)

Drag options to blanks, or click blank then click option'

Adictionary

Bcorpus

Cdocs

DLdaModel

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the corpus instead of the dictionary.

Passing the list of documents instead of the dictionary.

The id2word parameter expects the dictionary object that maps word IDs to words.

4fill in blank

hard

Fill both blanks to print the top 3 words for each topic in the LDA model.

NLP

for i in range([1]):
    print(f"Topic {i}:", lda.show_topic(i, topn=[2]))

Drag options to blanks, or click blank then click option'

A2

B3

C5

D10

Attempts:

3 left

💡 Hint

Common Mistakes

Using the wrong number of topics in the range.

Requesting more or fewer top words than needed.

We created 2 topics, so the range should be 2. We want the top 3 words, so topn=3.

5fill in blank

hard

Fill all three blanks to get the topic distribution for the first document and print the dominant topic index.

NLP

doc_bow = corpus[0]
topic_dist = lda.get_document_topics([1])
dominant_topic = max(topic_dist, key=lambda x: x[[2]])[[3]]
print(f"Dominant topic index: {dominant_topic}")

Drag options to blanks, or click blank then click option'

Adoc_bow

B1

C0

Dcorpus

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the whole corpus instead of a single document.

Mixing up the tuple indices for topic id and probability.

We pass the bag-of-words of the first document to get_document_topics. The topic distribution is a list of (topic_id, probability) tuples. To find the dominant topic, we use max with a key on the probability (index 1) and then get the topic id (index 0).