Machine Learning (ML) · Advanced level

Advanced level · Module 3 of 6 · 22 guided hours

Time Series and Forecasting

Forecast quantities over time with baselines, feature-based models and validation that respects time order.

Learning objectives

  • Describe trend, seasonality and noise in time series
  • Build forecasting baselines and feature-based models
  • Validate forecasts with backtesting and choose suitable error measures

Topics

  • trend, seasonality and noise
  • naive and seasonal-naive baselines
  • lag and rolling features
  • time-based backtesting
  • MAE and MAPE for forecasts
  • forecast uncertainty

Lessons

  1. Lesson 1 Understanding Time Series

    Recognise trend, seasonality and irregular movements, and prepare time series for modelling.

    trendseasonalitynoiseresampling

  2. Lesson 2 Baselines and Feature-Based Forecasting

    Start with naive baselines, then use lag and rolling features with regression models.

    naive forecastseasonal-naive forecastlag featuresrolling statistics

  3. Lesson 3 Backtesting Forecasts

    Validate forecasts by repeatedly training on the past and testing on the following period.

    backtestingrolling originTimeSeriesSplitforecast horizon

  4. Lesson 4 Forecast Accuracy and Uncertainty

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

    MAE and MAPEscale-free errorsprediction intervalscommunicating uncertainty

Practical activity

Forecast monthly sorghum and maize prices for four markets, compare models with seasonal baselines through backtesting and report accuracy per market.

Expected outcomes

  • Honest forecasts that beat baselines
  • Correct validation for time-ordered data

Assessment

Backtesting report marked on baselines, validation design and per-market interpretation.

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.