Challenges in Data Preparation CT-AI Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A team is building a model to predict whether a patient will be readmitted to a hospital within 30 days. The dataset includes features such as 'number_of_previous_admissions' and 'length_of_stay'. The team preprocesses the data by scaling all features using the mean and standard deviation computed from the entire dataset, and then splits the data into training and test sets. What is the main issue with this preprocessing approach?

   • AScaling is not necessary for tree-based models, so the preprocessing is wasted effort.
   • BThe scaling will cause the model to overfit to the training data because the test set is not scaled.
   • CScaling before splitting causes the test set to influence the training set, leading to data leakage.
   • DThe scaling will remove outliers, which are important for predicting readmission.
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A data scientist is cleaning a dataset for a regression model. The 'salary' feature has some extreme values that are much higher than the rest (e.g., a few values above $1,000,000 while most are below $100,000). The scientist wants to reduce the impact of these outliers. What is the most appropriate technique?

   • AApply a log transformation to the salary feature.
   • BReplace the outliers with the mean salary.
   • CUse min-max scaling to bring the values into a smaller range.
   • DRemove the outliers from the dataset.
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which of the following is an example of feature engineering?

   • ACreating a new feature 'age_in_years' from a 'birth_date' column
   • BRemoving duplicate rows
   • CSplitting data into train and test sets
   • DReplacing missing values with the median
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A data scientist is working on a binary classification problem with a highly imbalanced dataset. The scientist decides to use random undersampling of the majority class to balance the dataset. What is the primary disadvantage of this approach?

   • AIt can lead to loss of important information from the majority class, potentially reducing model performance.
   • BIt is computationally expensive because it requires generating synthetic samples.
   • CIt is only applicable to tree-based models.
   • DIt can cause the model to overfit to the minority class because of duplicate samples.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   Which data cleaning technique involves replacing missing values with a calculated statistic such as the mean or median of the column?

   • ANormalization
   • BOne-hot encoding
   • COutlier removal
   • DImputation
   Reveal Solution Discussion Select an answer first