Manage Data Transformation to Conduct Data Preparation PMI-CPMAI Practice Questions (Page 7)

Questions 31–35

1. 31application · medium

   A data science team is building a credit risk model with 500 features. They suspect many features are redundant and want to reduce dimensionality to improve model interpretability and reduce overfitting. They need a method that evaluates features based on their statistical relationship with the target variable, independent of any specific machine learning model. Which feature selection approach should they use?

   • APrincipal component analysis (PCA)
   • BWrapper method using recursive feature elimination
   • CFilter method using chi-squared test
   • DEmbedded method using Lasso regression
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   Which technique is an example of data normalization?

   • APrincipal component analysis
   • BOne-hot encoding
   • CZ-score standardization
   • DMin-max scaling
   Reveal Solution Discussion Select an answer first
4. 33foundation · easy

   What does data standardization do to a feature?

   • ACreates new features from the existing one
   • BScales it to a range between 0 and 1
   • CTransforms it to have zero mean and unit variance
   • DRemoves outliers from the feature
   Reveal Solution Discussion Select an answer first
5. 34application · medium

   A data science team is building a preprocessing pipeline for a dataset that contains both numerical and categorical features. The pipeline needs to handle missing values, scale numerical features, and encode categorical features. They want to ensure the pipeline is reusable and can be applied consistently to new data. Which approach should they use?

   • AUse a single function that performs all transformations at once without fitting
   • BWrite a script that applies each transformation manually to the training and test sets
   • CApply transformations in a random order each time
   • DCreate a scikit-learn Pipeline that chains imputation, scaling, and encoding steps
   Reveal Solution Discussion Select an answer first
6. 35foundation · easy

   What is the main goal of feature selection methods?

   • ATo choose the most relevant features while reducing dimensionality
   • BTo generate synthetic samples for training
   • CTo create new features from existing ones
   • DTo transform data to have zero mean and unit variance
   Reveal Solution Discussion Select an answer first