Supervised Learning DY0-001 Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   In a linear regression model y = β0 + β1*x, what does β1 represent?

   • AThe slope of the line
   • BThe predicted value of y
   • CThe error term
   • DThe intercept of the line
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   How can k-nearest neighbors (KNN) be adapted for a regression task?

   • ABy taking the average of the target values of the k nearest neighbors
   • BBy taking the majority vote of the k nearest neighbors
   • CBy selecting the most frequent target value
   • DBy fitting a logistic function to the neighbors
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A hospital wants to predict whether a patient will be readmitted within 30 days (binary outcome). The dataset includes age, blood pressure, number of prior visits, and a categorical 'insurance type' field. The team wants to understand the impact of each feature on the log-odds of readmission. Which model is most appropriate, and what is a key consideration for the categorical feature?

   • ALogistic regression; one-hot encode insurance type
   • BNaive Bayes; assume insurance type is conditionally independent
   • CK-nearest neighbors; use raw categorical values directly
   • DLinear regression; encode insurance type as ordinal integers
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A marketing team wants to classify new customers into one of three segments (budget, standard, premium) based on their purchase history. The dataset has 50,000 customers with 20 numeric features. The team needs a model that can capture non-linear decision boundaries and requires no training phase (they want to add new customers instantly without retraining). Which algorithm best fits these constraints?

   • ALinear regression
   • BNaive Bayes
   • CK-nearest neighbors with k=5
   • DLogistic regression (one-vs-rest)
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A data analyst is using KNN to classify images of handwritten digits (28x28 pixels, 784 features). The model performs poorly, and the analyst suspects the high dimensionality is hurting the distance metric. Which action is most likely to improve KNN performance in this scenario?

   • ANormalize the pixel values to a range of 0-1
   • BIncrease k to 50
   • CUse Manhattan distance instead of Euclidean
   • DApply principal component analysis (PCA) to reduce dimensions before KNN
   Reveal Solution Discussion Select an answer first