K-NN algorithm for machine learning

What is kNN?

• A clustering algorithm

• A reinforcement learning algorithm

• A supervised learning algorithm

• An unsupervised learning algorithm

How does kNN classify new examples?

• By finding the most dissimilar training examples

• By choosing the k closest neighbors

• By ignoring all training examples

• Based on random selection

What is the purpose of scaling in kNN?

• To reduce the number of attributes

• To remove irrelevant attributes

• To make the data more complex

• To normalize the variables to the same scale

What is the distance formula used in kNN?

• Cosine Similarity

• Chebyshev Distance

• Minkowski Distance

• Manhattan Distance

How does kNN handle noise in labels?

• By assigning a random label

• By increasing the value of k

• By choosing the most frequent label of its k nearest neighbors

• By discarding the noisy examples

What is an advantage of kNN?

• Only suitable for regression problems

• Performs poorly with large datasets

• Simple to implement algorithm

• Requires extensive tuning

What is the effect of increasing k in kNN?

• Has no impact on the algorithm

• Reduces the number of neighbors considered

• Increases the possibility of errors

• Decreases sensitivity to noise

Which scaling technique rescales features using the median and quartile range?

• Min-Max Scaling

• Normalization

• Standardization

• Robust Scaling

What is the purpose of the Curse-of-Dimensionality in kNN?

• To improve prediction accuracy

• To reduce the number of attributes

• To prevent overfitting

• To highlight the importance of relevant attributes

How does kNN handle prediction for new examples?

• By predicting the same output as the farthest neighbor

• By ignoring all training examples

• By predicting the average of all neighbors

• By predicting the most frequent label of its k nearest neighbors