Forecasting Vocabulary

Vocabulary flashcards covering key terms, principles, formulas, and accuracy measures in forecasting.

Independent Demand

Demand for a finished product or item that comes directly from external sources, such as customer orders, and is not dependent on the demand for other items.

Dependent Demand

Demand for raw materials, component parts, or subassemblies that depends directly on the demand for a higher-level item or finished product.

Qualitative Forecasting Techniques

Forecasting methods based on subjective estimates, expert opinion, and qualitative data, which include historical analogy, marketing research, management estimation, build-up forecast, panel consensus, and the Delphi method.

Time Series Forecasting

A quantitative forecasting technique that uses chronologically ordered historical data to predict future values.

Naïve Forecast

A quantitative forecasting method that assumes demand in the next period will equal the actual demand in the most recent period ($F_t = D_{t-1}$).

Simple Moving Average

An arithmetic average of a specified number ($n$) of the most recent historical observations, where the oldest observation is dropped as each new observation is added ($F_t = \frac{D_{t-1} + D_{t-2} + D_{t-3} + \times \times \times + D_{t-n}}{n}$).

Weighted Moving Average

An averaging technique in which historical observations are assigned specific weights ($w_t$) according to their importance, where all weights must sum to 1 ($\frac{\frac{n}{i=1}}{} w_i = 1$).

Exponential Smoothing

A type of weighted moving average forecasting technique in which past observations are geometrically discounted according to their age using an alpha smoothing constant (\frac{\frac{\frac{\frac{\frac{0}{1}}{\times}}{\times}}{\times}}{\times} \frac{\frac{\frac{\frac{1}{1}}{\times}}{\times}}{\times}}{\times}), computed as F_t = F_{t-1} + \frac{\frac{\frac{\frac{\frac{0}{1}}{\times}}{\times}}{\times}}{\times} \frac{\frac{\frac{\frac{1}{1}}{\times}}{\times}}{\times}}{\times}(D_{t-1} - F_{t-1}).

Forecast Error

The difference between actual demand and forecast value for a given period ($FE_t = D_t - F_t$).

Forecast Bias

A consistent error trend caused by issues in the forecasting technique (such as trend change lag) or human optimism/pessimism that leads to consistent overforecasting or underforecasting.

Common Causes

Numerous uncontrollable, everyday factors that naturally affect every process and create routine forecast variation.

Mean Forecast Error (MFE)

The average of forecast errors over $n$ periods (MFE = \frac{\frac{\frac{\frac{\frac{\frac{\frac{\frac{FE}{n}}{1}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times} = \frac{\frac{\frac{\frac{\frac{\frac{\frac{(\text{Actual} - \text{Forecast})}{\text{Number of Periods}}}{1}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}), where a negative value indicates overforecasting and a positive value indicates underforecasting.

Mean Absolute Deviation (MAD)

A measure of average forecast error magnitude calculated by averaging the absolute values of all forecast errors (MAD = \frac{\frac{\frac{\frac{\frac{\frac{\frac{|FE_t|}{n}}{1}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}), where an ideal value of zero indicates no forecasting error.

Mean Absolute Percentage Error (MAPE)

A forecast accuracy measure that provides the perspective of the true relative magnitude of forecast error expressed as a percentage (MAPE = \frac{1}{n} \frac{\frac{\frac{\frac{\frac{\frac{\frac{n}{t=1}}{1}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times}}{\times} \times |\frac{FE_t}{D_t}|(100)).

Tracking Signal (TS)

A metric that indicates whether the forecast average is keeping pace with genuine upward or downward changes in demand, calculated as $TS = \frac{\text{Algebraic Sum of Forecast Errors}}{\text{Mean Absolute Deviation}}$.