Objective 2.1 Recognize Relative Impact of Data Quality and Size to Algorithms AIP-210 Practice Questions (Page 3)

Questions 11–15

1. 11application · medium

   A data science team is building a recommendation system. They have two datasets: Dataset A has 5 million user-item interactions but is known to contain many bot-generated fake interactions. Dataset B has 200,000 genuine interactions from real users. They have limited time to train and evaluate models. Which dataset should they use as the primary training set?

   • ADataset A, but only after removing all bot interactions, even if that leaves fewer than 200,000 records
   • BDataset B, because smaller datasets always train faster and are easier to tune
   • CDataset B, because the fake interactions in Dataset A will introduce systematic bias that more data cannot correct
   • DDataset A, because the sheer volume will help the model generalize better
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A team is training a regression model to predict housing prices. They notice that a small number of records have sale prices that are 10 times higher than the median, and these records also have unusually large square footage values. The model's performance on a validation set is poor, especially for typical homes. What is the most likely cause of this degradation?

   • AThe model is underfitting because the dataset is too small
   • BThe features are not normalized, causing gradient descent to diverge
   • CThe target variable is skewed, so the model predicts the mean for all inputs
   • DOutliers in the training data are skewing the model's learned relationship
   Reveal Solution Discussion Select an answer first
4. 13expert · hard

   A data scientist is assessing the quality of a dataset for a predictive maintenance model. They run a profiling script and find that the 'temperature' feature has a mean of 75°F, a standard deviation of 2°F, and a maximum of 212°F. The 'vibration' feature has a mean of 0.5, a standard deviation of 0.1, and a maximum of 100. What should they conclude from these statistics?

   • AThe vibration feature is complete because it has a non-zero standard deviation
   • BThe temperature feature is accurate because its mean is reasonable
   • CThe temperature feature is more reliable than the vibration feature because its standard deviation is smaller
   • DBoth features likely contain outliers that could distort the model
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   Which data quality dimension is most directly concerned with whether the data reflects the true real-world values it is supposed to represent?

   • AAccuracy
   • BCompleteness
   • CTimeliness
   • DConsistency
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A data scientist is exploring a dataset for a binary classification problem. They create a histogram of the target variable and see that 95% of the samples belong to class 0 and only 5% to class 1. What does this visualization most directly reveal?

   • AMissing values in the target variable
   • BThe presence of outliers in the feature space
   • CThe correlation between features and the target
   • DA class imbalance that could bias the model toward the majority class
   Reveal Solution Discussion Select an answer first