Objective 3.3 Carry Out Feature Engineering CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A data scientist is working with a categorical feature 'UserAgent' that has over 1 million unique values in a large dataset. The model is a linear classifier, and the data scientist needs to handle this feature without creating an enormous feature space. Which encoding method is most appropriate?

   • ATarget encoding
   • BBinary encoding
   • COne-hot encoding
   • DHash encoding with a fixed number of columns
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   A categorical feature has 100 unique categories. The data scientist wants to reduce dimensionality compared to one-hot encoding while still capturing the category information. Which encoding method should they use?

   • ATarget encoding
   • BOne-hot encoding
   • CBinary encoding
   • DLabel encoding
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A data scientist is working with a sparse matrix of user-item interactions with 1 million rows and 10,000 columns. The goal is to reduce dimensionality for a recommendation system. Which method is most appropriate?

   • APCA
   • BFactor analysis
   • CSingular Value Decomposition (SVD)
   • Dt-SNE
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   What is the effect of applying the reverse operation to the string 'data science'?

   • AIt becomes 'science data'.
   • BIt becomes 'data'.
   • CIt becomes 'ecneics atad'.
   • DIt becomes 'DATA SCIENCE'.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   A dataset contains a column 'full_name' with values like 'John Smith'. The data scientist wants to create separate features for the first and last names. Which feature engineering technique should they use?

   • ASplit features
   • BTarget encoding
   • COne-hot encoding
   • DPCA
   Reveal Solution Discussion Select an answer first