Define Machine Learning Concepts for Data Science Workloads SNOWPRO-ADVANCED-DATA-SCIENTIST Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   A data scientist notices that a model's performance improves after removing several highly correlated features. Which feature engineering technique does this illustrate?

   • AFeature extraction
   • BFeature selection
   • CFeature scaling
   • DData augmentation
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   Which evaluation metric is most appropriate for a binary classification problem where the positive class is rare (e.g., detecting fraud in transactions)?

   • AR-squared
   • BPrecision and recall
   • CMean squared error
   • DAccuracy
   Reveal Solution Discussion Select an answer first
4. 18expert · hard

   A data scientist is building a model to predict customer lifetime value. They have a feature that is highly correlated with the target, but it is only available for 50% of the records. Which of the following is the best approach?

   • ADrop the feature entirely because it has too many missing values.
   • BImpute the missing values using a model trained on the available data.
   • CReplace missing values with the mean of the feature.
   • DDrop all rows with missing values.
   Reveal Solution Discussion Select an answer first
5. 19foundation · easy

   Which algorithm is best suited for predicting a continuous numerical value, such as house price, based on features like square footage and location?

   • ALogistic regression
   • BDecision tree for classification
   • CK-means clustering
   • DLinear regression
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A data scientist is preparing a dataset for model training. To evaluate the model's generalization, which of the following is the most appropriate approach?

   • ASplit the data into training, validation, and test sets.
   • BRandomly shuffle the data and pick the best features after evaluation.
   • CUse only a training and test split, without validation.
   • DTrain on all data and evaluate on the same data.
   Reveal Solution Discussion Select an answer first