Feature engineering for numerical and categorical variables: Practice Questions — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)

Practice Questions: Feature Engineering for Numerical and Categorical Variables

Feature engineering is a critical skill tested in the NVIDIA-Certified Associate: Accelerated Data Science exam, especially for numerical and categorical variables. Below are multiple-choice questions designed to help you prepare effectively.

1. Which technique is most appropriate for converting a categorical variable with no intrinsic order into a format suitable for machine learning models?

   • A) Label Encoding
   • B) One-Hot Encoding
   • C) Min-Max Scaling
   • D) Principal Component Analysis (PCA)

   Correct Answer: B) One-Hot Encoding

   Explanation: One-Hot Encoding creates binary columns for each category, preserving the nominal nature without implying order, unlike Label Encoding which assigns arbitrary numeric values.

2. When dealing with a numerical feature that has a highly skewed distribution, which transformation is commonly applied to improve model performance?

   • A) Logarithmic Transformation
   • B) One-Hot Encoding
   • C) Label Encoding
   • D) Synthetic Minority Over-sampling Technique (SMOTE)

   Correct Answer: A) Logarithmic Transformation

   Explanation: Log transformation reduces skewness by compressing large values, making the distribution more normal-like, which benefits many algorithms.

3. Which method is best suited for encoding an ordinal categorical variable, such as education level (e.g., High School, Bachelor, Master, PhD)?

   • A) One-Hot Encoding
   • B) Label Encoding
   • C) Binning
   • D) Min-Max Scaling

   Correct Answer: B) Label Encoding

   Explanation: Label Encoding assigns integer values that reflect the order inherent in ordinal variables, preserving their ranking.

4. In the context of GPU-accelerated data science, which library is optimized for efficient feature engineering on large datasets with categorical variables?

   • A) pandas
   • B) cuDF
   • C) NumPy
   • D) Matplotlib

   Correct Answer: B) cuDF

   Explanation: cuDF is a GPU-accelerated DataFrame library designed to handle large-scale data manipulation and feature engineering efficiently on NVIDIA GPUs.

5. Which feature engineering technique helps reduce dimensionality when dealing with high-cardinality categorical variables?

   • A) Frequency Encoding
   • B) One-Hot Encoding
   • C) Label Encoding
   • D) Synthetic Data Generation

   Correct Answer: A) Frequency Encoding

   Explanation: Frequency Encoding replaces categories with their frequency counts, reducing the number of features compared to One-Hot Encoding, which can create many sparse columns.

6. What is the primary purpose of binning numerical variables during feature engineering?

   • A) To convert numerical data into categorical data
   • B) To increase the number of features
   • C) To normalize the data
   • D) To generate synthetic samples

   Correct Answer: A) To convert numerical data into categorical data

   Explanation: Binning groups continuous numerical values into discrete intervals, which can help capture nonlinear relationships and reduce noise.

7. Which technique is effective for handling missing values in categorical variables before feature encoding?

   • A) Imputing with the mode
   • B) Dropping the entire feature
   • C) Logarithmic Transformation
   • D) Min-Max Scaling

   Correct Answer: A) Imputing with the mode

   Explanation: Imputing missing categorical data with the mode (most frequent category) maintains data integrity and prepares the feature for encoding.

8. Why is feature scaling generally not applied to categorical variables?

   • A) Because categorical variables have no numeric meaning
   • B) Because scaling is computationally expensive
   • C) Because it improves model accuracy
   • D) Because it reduces dimensionality

   Correct Answer: A) Because categorical variables have no numeric meaning

   Explanation: Scaling applies to numerical features to normalize ranges; categorical variables represent discrete classes without inherent numeric scale, so scaling is inappropriate.

More in this topic

Data integration and manipulation with cuDF and pandas: Quick Reference — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Feature engineering for numerical and categorical variables — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Dimensionality reduction and data sampling: Worked Example — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data integration and manipulation with cuDF and pandas: Worked Example — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Efficient processing and storage with Parquet — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Feature engineering for numerical and categorical variables: Common Mistakes — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data integration and manipulation with cuDF and pandas: Practice Questions — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data cleaning, quality handling, and governance: Worked Example — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data cleaning, quality handling, and governance: Common Mistakes — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data cleaning, quality handling, and governance: Practice Questions — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data integration and manipulation with cuDF and pandas — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data integration and manipulation with cuDF and pandas: Common Mistakes — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Dimensionality reduction and data sampling: Practice Questions — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data cleaning, quality handling, and governance — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data cleaning, quality handling, and governance: Quick Reference — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)GPU-accelerated ETL with RAPIDS, Dask, or Spark — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Feature engineering for numerical and categorical variables: Quick Reference — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Dimensionality reduction and data sampling — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Feature engineering for numerical and categorical variables: Worked Example — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Dimensionality reduction and data sampling: Quick Reference — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Dimensionality reduction and data sampling: Common Mistakes — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Handling class imbalance and generating synthetic data — Data Manipulation and Preparation (NVIDIA-Certified Associate: Accelerated Data Science)Data Manipulation and Preparation — NVIDIA-Certified Associate: Accelerated Data Science