Supervised Learning - Prediction Week 3 Challenge

2.

L1 and L2 gradient descent methods differ in their regularization techniques. L1 uses the absolute values of coefficients (Lasso), promoting sparsity, while L2 uses squared values (Ridge), penalizing large coefficients. Lasso helps for feature selection as well.

3.

Different loss functions for regression include Mean Squared Error (MSE), Mean Absolute Error (MAE), Huber loss, and Quantile loss, each with unique properties suitable for different scenarios.

4.

Learning rate determines the step size in gradient descent, influencing convergence speed and stability. It balances between rapid convergence and overshooting, impacting the optimization process significantly.

5.

Linear regression is evaluated using metrics like Mean Squared Error (MSE), Mean Absolute Error (MAE), R-squared, and Adjusted R-squared, assessing how well the model fits the data and its predictive performance.

6.

Multiple coefficient of determination (R-squared) measures the proportion of variance explained by all predictors, while adjusted R-squared adjusts for the number of predictors, penalizing for additional terms, providing a more accurate evaluation of the model's goodness of fit.