Objective 2.4 Apply Problem-Specific Transformations to Data Sets CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   In TF-IDF, why is the inverse document frequency (IDF) component important?

   • AIt increases the weight of common words like 'the' and 'and'
   • BIt reduces the weight of words that appear in many documents, highlighting rare words
   • CIt converts the term frequency into a binary value
   • DIt normalizes the term frequency by the document length
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   What is the purpose of generating latent representations for image data?

   • ATo create a compact, meaningful feature vector that captures the essential information of an image
   • BTo label each pixel with a semantic category
   • CTo convert an image into a binary format for storage
   • DTo increase the resolution of an image for better display
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which of the following techniques can be used to generate latent representations for images?

   • ATF-IDF
   • BOne-hot encoding
   • CAutoencoders
   • DWord2Vec
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A data scientist is building a sentiment analysis model for customer reviews. The reviews contain emojis, hashtags, and URLs. The team wants to tokenize the text so that emojis are treated as separate tokens, hashtags are split into words (e.g., '#great' becomes 'great'), and URLs are removed. Which tokenization strategy should they use?

   • AUse a simple whitespace tokenizer.
   • BUse a pre-trained subword tokenizer (e.g., BPE) without any preprocessing.
   • CUse a regular expression-based tokenizer that matches words, emojis, and hashtags, and filters out URLs.
   • DUse a tokenizer that only splits on punctuation.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A data science team is building a text classification model for customer feedback. They have a large corpus of feedback and want to represent each feedback as a vector that captures the semantic meaning of the words, not just their presence. They also want the representation to be dense and low-dimensional. Which vectorization method should they use?

   • AOne-hot encoding of the feedback text.
   • BBag-of-words (count vectorizer).
   • CAveraged Word2Vec embeddings of the feedback text.
   • DTF-IDF vectorization.
   Reveal Solution Discussion Select an answer first