Objective 4.2 Train Models CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 8)

Questions 36–40

1. 36foundation · easy

   What type of data is an ARIMA model designed to forecast?

   • AText data for sentiment analysis
   • BTime series data
   • CUnlabeled data for clustering
   • DCross-sectional data
   Reveal Solution Discussion Select an answer first
2. 37foundation · easy

   Which theorem forms the basis of the naïve Bayes classifier?

   • ACentral limit theorem
   • BBayes' theorem
   • CFermat's little theorem
   • DPythagorean theorem
   Reveal Solution Discussion Select an answer first
4. 38application · medium

   A data scientist is training a support vector machine (SVM) classifier and needs to tune the regularization parameter C and the kernel coefficient gamma. They want to find the best combination efficiently. Which approach is most systematic?

   • AManual tuning based on intuition
   • BUse default hyperparameters without tuning
   • CRandom search with a single train-test split
   • DGrid search with cross-validation
   Reveal Solution Discussion Select an answer first
5. 39foundation · easy

   What is a key advantage of density-based clustering like DBSCAN over k-means?

   • AIt is faster on large datasets
   • BIt always produces the same result regardless of parameters
   • CIt requires the number of clusters to be specified in advance
   • DIt can find clusters of arbitrary shape
   Reveal Solution Discussion Select an answer first
6. 40application · medium

   A manufacturing company wants to predict the daily output of a production line based on temperature, humidity, machine speed, and shift. The data contains non-linear relationships and some missing values. The team needs a model that can handle these characteristics without extensive data preprocessing and wants to assess its generalization reliably. Which approach should they use?

   • ALogistic regression with k-fold cross-validation
   • BARIMA with a single train-test split
   • CRandom forest regression with k-fold cross-validation
   • DLinear regression with a simple train-test split
   Reveal Solution Discussion Select an answer first