Regression

Question 1

Y = f(X) + e

Answer 1

Y = numeric output
e = error term
f = systematic information X provides about Y

Goal: find an f^(x) that approximates the true function f(x) as well as possible

Question 2

MSE

Answer 2

Represents averaged squared difference of a prediction ^y = ^f(x) from its true value y.

MSE= E[(y-6f(x))^2]

Question 3

How to minimize reducible error

Answer 3

Using a more sophisticated or appropriate model.

Collecting more relevant data or improving data quality.

Tuning model hyperparameters for better performance.

Question 4

Bias

Answer 4

High bias can cause an algorithm to miss the relevant relations between features and target outputs (underfitting)

Question 5

Variance

Answer 5

High variance may result from an algorithm modeling the random noise in the training data (overfitting)