Machine Learning (ML) · Advanced level

Advanced level · Module 3: Time Series and Forecasting · Lesson 4 of 4

Forecast Accuracy and Uncertainty

Measure forecast error appropriately and communicate uncertainty to decision-makers.

In this lesson you will

  • Choose between MAE, RMSE and MAPE
  • Provide prediction intervals and communicate them

Key concepts

MAE and MAPEscale-free errorsprediction intervalscommunicating uncertainty

The lesson

MAE in currency units is easy to explain; MAPE expresses error as a percentage, which helps compare markets with different price levels but breaks down when values are near zero.

Point forecasts hide uncertainty. Prediction intervals — for example from quantile regression or past error distributions — give a range such as '5,200 to 6,100 SSP with 80% confidence'.

Tell decision-makers how often actual values fell inside the interval in backtests, and remind them that shocks outside past experience can break any forecast.

Step by step

  1. Calculate MAE and MAPE for each market.
  2. Build 80% prediction intervals.
  3. Check interval coverage in the backtest.
  4. Write a forecast statement with its range.

Practice activity

Write the forecast summary for the four markets with accuracy and uncertainty.

Expected result: Per-market accuracy and calibrated intervals explained in plain language.

Check your understanding

  1. When does MAPE become unreliable?

    Show answer

    When actual values are close to zero.

  2. What does interval coverage measure?

    Show answer

    How often actual values fell inside the prediction intervals.

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.