Machine Learning (ML) · Beginner level

Beginner level · Module 4: Evaluating Machine Learning Models · Lesson 1 of 3

Measuring Regression Errors

Measure how far numeric predictions are from the truth with MAE, RMSE and R².

In this lesson you will

  • Calculate and interpret MAE and RMSE
  • Explain what R² does and does not say

Key concepts

mean absolute errorroot mean squared errorR²error in real units

The lesson

Mean absolute error (MAE) is the average size of the errors, in the same units as the target: an MAE of 120 kg per hectare is easy to explain. Root mean squared error (RMSE) squares errors before averaging, so large mistakes count more.

R² compares the model with always predicting the average: 1 is perfect, 0 is no better than the average, and it can even be negative. A high R² does not mean predictions are accurate enough for the decision.

Always relate error to the decision: is an average error of 120 kg/ha small or large for a farmer planning sales?

Step by step

  1. Calculate MAE by hand for five predictions.
  2. Calculate MAE, RMSE and R² with scikit-learn.
  3. Compare the regression model with the mean baseline.
  4. Explain the results in plain language.

Practice activity

Evaluate the crop-yield model and write two sentences a farmer could understand about its accuracy.

Expected result: Correct metrics and a plain-language explanation of typical error.

Check your understanding

  1. Which metric punishes large errors more, MAE or RMSE?

    Show answer

    RMSE.

  2. What does an R² of 0 mean?

    Show answer

    The model is no better than always predicting the average.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.