Topic coherence evaluation in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to calculate the coherence score using Gensim's CoherenceModel.

NLP

from gensim.models.coherencemodel import CoherenceModel
coherence_model = CoherenceModel(model=lda_model, texts=tokenized_texts, dictionary=dictionary, coherence='[1]')
coherence_score = coherence_model.get_coherence()

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

Ac_v

Bc_uci

Cc_npmi

Du_mass

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'u_mass' which requires a corpus in BoW format and is less interpretable.

Confusing coherence types and using one not supported by the model.

The 'c_v' coherence measure is commonly used for topic coherence evaluation as it combines indirect cosine similarity with a sliding window and a boolean sliding window.

2fill in blank

medium

Complete the code to preprocess texts by tokenizing and removing stopwords before coherence evaluation.

NLP

from nltk.corpus import stopwords
stop_words = set(stopwords.words('english'))
processed_texts = [[word for word in doc.lower().split() if word not in [1]] for doc in documents]

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

Atexts

Bdictionary

Ctokenized_texts

Dstop_words

Attempts:

3 left

💡 Hint

Common Mistakes

Using the dictionary object instead of stopwords set.

Not converting words to lowercase before filtering.

Stopwords are common words to remove before analysis. The variable 'stop_words' holds the set of stopwords to filter out.

3fill in blank

hard

Fix the error in the code to compute coherence score by correctly passing the dictionary parameter.

NLP

coherence_model = CoherenceModel(model=lda_model, texts=tokenized_texts, [1]=dictionary, coherence='c_v')
score = coherence_model.get_coherence()

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

Adictionary

Bdict

Ccorpus

Dtokens

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'dict' which is a Python keyword, not a parameter name.

Passing the corpus instead of dictionary.

The parameter name for the dictionary in CoherenceModel is 'dictionary'. Using 'dict' or other names causes errors.

4fill in blank

hard

Fill both blanks to create a dictionary and corpus needed for topic coherence evaluation.

NLP

from gensim import corpora
[1] = corpora.Dictionary(tokenized_texts)
[2] = [[1].doc2bow(text) for text in tokenized_texts]

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

Adictionary

Bcorpus

Ctexts

Dtokens

Attempts:

3 left

💡 Hint

Common Mistakes

Swapping dictionary and corpus variable names.

Using undefined variable names.

The dictionary maps words to ids, and the corpus is a list of bag-of-words representations of documents.

5fill in blank

hard

Fill all three blanks to compute and print the coherence score for an LDA model.

NLP

coherence_model = CoherenceModel(model=[1], texts=[2], dictionary=[3], coherence='c_v')
score = coherence_model.get_coherence()
print(f"Coherence Score: {score:.4f}")

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

Alda_model

Btokenized_texts

Cdictionary

Dcorpus

Attempts:

3 left

💡 Hint

Common Mistakes

Using corpus instead of tokenized texts for the texts parameter.

Passing wrong variable names causing runtime errors.

The LDA model, tokenized texts, and dictionary are required to compute the coherence score.