Objective 2.4 Transform Numerical and Categorical Data AIP-210 Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Which categorical encoding technique creates a binary column for each category, where exactly one column is 1 for each observation, and is best suited for nominal categories with no inherent order?

   • AOne-hot encoding
   • BLabel encoding
   • CFrequency encoding
   • DOrdinal encoding
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which strategy is most appropriate for handling missing values in a numerical feature that is approximately normally distributed and has a small percentage of missing data?

   • AImpute with the median
   • BImpute with the mode
   • CImpute with the mean
   • DDelete all rows with missing values
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which numerical transformation technique rescales data to a fixed range, typically [0, 1], and is most appropriate when the distribution is not Gaussian and the algorithm does not assume a specific distribution?

   • ARobust scaling
   • BStandardization (z-score)
   • CMin-max scaling
   • DLog transformation
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   A categorical feature has over 100 unique categories. Which encoding technique is most appropriate to avoid creating an excessive number of new columns?

   • AOne-hot encoding
   • BFrequency encoding
   • CBinary encoding
   • DOrdinal encoding
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A data scientist is building a linear regression model to predict energy consumption. The 'Temperature' feature has a few extreme values (e.g., -20°C and 45°C) that are valid weather events. The model is sensitive to outliers, and the team wants to reduce their influence while preserving the true relationship between temperature and consumption. The data is not normally distributed. Which approach should the data scientist take?

   • AApply a log transformation to the Temperature feature.
   • BApply robust scaling using the median and IQR.
   • CApply winsorization to cap the Temperature values at the 5th and 95th percentiles.
   • DRemove the rows with extreme Temperature values.
   Reveal Solution Discussion Select an answer first