Bird
0
0

What will happen if you run t-SNE with perplexity set too high on a small dataset?

medium📝 Predict Output Q5 of 15
NLP - Word Embeddings
What will happen if you run t-SNE with perplexity set too high on a small dataset?
AThe output will have more than 3 dimensions
Bt-SNE will run faster and produce better clusters
Ct-SNE may produce poor or unstable visualizations
Dt-SNE will ignore the perplexity parameter
Step-by-Step Solution
Solution:

  1. Step 1: Understand perplexity role

    Perplexity controls balance between local and global aspects; too high on small data causes issues.

  2. Step 2: Effects of high perplexity on small data

    It leads to poor or unstable visualizations because the algorithm tries to consider too many neighbors.

  3. Final Answer:

    t-SNE may produce poor or unstable visualizations -> Option C

  4. Quick Check:

    High perplexity + small data = poor visualization [OK]
Quick Trick: Keep perplexity smaller than dataset size [OK]
Common Mistakes:
MISTAKES
  • Assuming higher perplexity always improves results
  • Expecting more output dimensions
  • Thinking parameter is ignored

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