KNN

Question 1

what is the KNN prediction algorithm?

Answer 1

1. Identify the K training with the smallest distance to your test point
2. determine the conditional probability of each class within your “neighborhood”
3. set your test point to the class with the higheset probability

Question 2

what is a decision boundary?

Answer 2

its’ where the KNN algorithm will change it’s answer.

can sometimes misclassify since it classifies based off probabilities

Question 3

what is the relationship betweenthe choice of K and the model flexibility?

Answer 3

the lower the K, the higher the flexibility

Question 4

whats the best way to evaluate model accuracy?

Answer 4

percent of correctly classifed predictions

for error rate, percent of missclassified predictions