Machine Learning
Supervised Learning: Regression / Model Evaluation Metrics
Model Evaluation Metrics
How do we know if our regression model is performing well? We use evaluation metrics that calculate the error between predicted values (ŷ) and actual values (y).
Mean Absolute Error (MAE)
MAE = (1/n) * Σ|yi - ŷi|
Calculates the average absolute distance between predictions and actual values. It is more robust to outliers than MSE.
Mean Squared Error (MSE)
MSE = (1/n) * Σ(yi - ŷi)²
Penalizes larger errors more heavily by squaring the difference.
Root Mean Squared Error (RMSE)
RMSE = √MSE
Brings the error metric back to the original units of the target variable.
R-Squared (Coefficient of Determination)
R² = 1 - (SSR / SST)
Ranges from 0 to 1. Measures the proportion of variance in the target variable explained by the model.
Adjusted R-Squared
Adj R² = 1 - [(1 - R²) * (n - 1) / (n - k - 1)]
Unlike R², Adjusted R² only increases if a new feature improves the model more than would be expected by chance. (n = sample size, k = number of predictors)
Ready to test your Model Evaluation Metrics knowledge?
Model Evaluation Metrics
Quiz on understanding and applying regression metrics like MAE, MSE, RMSE, and R-Squared.