A measurement of whether the forecast error is consistently running higher or lower than actual demand.
Long-term variations influenced by macroeconomic factors and business cycles, often lasting several years (e.g., housing market shifts). A measurement of whether the forecast error is
The average of the squared differences between forecasted and actual values. Older editions relied heavily on classical statistics
Older editions relied heavily on classical statistics. The current version introduces how Machine Learning (ML) algorithms—such as Random Forests and Neural Networks—are handling complex, high-dimensional data sets that traditional regression cannot manage. Time series look inward at history; causal models
Measures the average magnitude of the mistakes regardless of direction, proving highly useful for determining safety stock levels.
Time series look inward at history; causal models look outward at external factors. The 3rd Edition expands on regression analysis, teaching planners how to correlate demand with variables like: