Bird
0
0

After running t-SNE, your plot shows all points clustered tightly with no clear groups. What debugging step should you try first?

medium📝 Debug Q7 of 15
NLP - Word Embeddings
After running t-SNE, your plot shows all points clustered tightly with no clear groups. What debugging step should you try first?
AIncrease the number of output dimensions to 10
BAdjust the perplexity parameter to a smaller value
CNormalize the input embeddings to zero mean and unit variance
DUse raw text instead of embeddings as input
Step-by-Step Solution
Solution:

  1. Step 1: Understand tight clustering cause

    Too high perplexity can cause points to cluster too tightly without separation.

  2. Step 2: Choose best debugging action

    Lowering perplexity often improves separation; other options are less relevant or incorrect.

  3. Final Answer:

    Adjust the perplexity parameter to a smaller value -> Option B

  4. Quick Check:

    Tight clusters fix = lower perplexity [OK]
Quick Trick: Tune perplexity to improve cluster separation [OK]
Common Mistakes:
MISTAKES
  • Increasing output dims unnecessarily
  • Feeding raw text instead of vectors
  • Skipping parameter tuning

Want More Practice?

15+ quiz questions · All difficulty levels · Free

Free Signup - Practice All Questions
More NLP Quizzes
Sentiment Analysis Advanced - Fine-grained sentiment (5-class) - Quiz 7medium Sentiment Analysis Advanced - Sentiment with context (sarcasm, negation) - Quiz 6medium Sequence Models for NLP - Embedding layer usage - Quiz 9hard Sequence Models for NLP - Why sequence models understand word order - Quiz 1easy Sequence Models for NLP - RNN for text classification - Quiz 9hard Text Similarity and Search - Jaccard similarity - Quiz 4medium Topic Modeling - Why topic modeling discovers themes - Quiz 14medium Word Embeddings - Word similarity and analogies - Quiz 8hard Word Embeddings - Pre-trained embedding usage - Quiz 15hard Word Embeddings - Pre-trained embedding usage - Quiz 9hard