Operations Management – Forecasting (Chapter 3)

These vocabulary flashcards cover key terms, concepts, metrics, models, and strategic considerations presented in Chapter 3 on forecasting in Operations Management. Reviewing them will strengthen understanding of both qualitative and quantitative forecasting techniques and their applications.

Forecast

A statement about the future value of a variable of interest, such as demand, weather, or resource availability.

Expected Level of Demand

The most likely quantity that will be required during a specified future period, often influenced by trend or seasonal patterns.

Forecast Accuracy

The closeness of forecasted values to actual outcomes; related to potential size of forecast error.

Plan the System

Long-range planning decisions on product mix, facility size, equipment levels, and facility location that rely on forecasts.

Plan the Use of the System

Short- and medium-range decisions—inventory, workforce, purchasing, production, budgeting, scheduling—guided by forecasts.

Features Common to All Forecasts

1) Assume past causal system persists, 2) Are never perfect, 3) Are more accurate for groups than individuals, 4) Lose accuracy as horizon increases.

Elements of a Good Forecast

Timely, accurate, reliable, written, in meaningful units, simple to understand, and cost-effective.

Steps in the Forecasting Process

1) Determine purpose, 2) Set time horizon, 3) Obtain & clean data, 4) Select technique, 5) Make forecast, 6) Monitor errors.

Forecast Error

Difference between actual value and forecast (Error = Actual − Forecast).

Mean Absolute Deviation (MAD)

Average of absolute forecast errors; weights all errors equally.

Mean Squared Error (MSE)

Average of squared forecast errors; penalizes larger errors more heavily.

Mean Absolute Percent Error (MAPE)

Average of absolute errors expressed as a percentage of actual values; measures relative error.

Qualitative Forecasting

Methods using subjective inputs—human judgment, expert opinions, consumer surveys—when hard data are scarce.

Quantitative Forecasting

Techniques that rely on numerical data, including time-series projections and associative (causal) models.

Executive Opinions

Forecasts developed collectively by a small group of upper-level managers based on strategic insight.

Sales Force Opinions

Forecast inputs from sales or customer-service staff who have direct customer contact and knowledge of future plans.

Consumer Surveys

Questionnaires or interviews that solicit customers’ purchase intentions to estimate future demand.

Time-Series Forecasting

Projection of patterns found in historical, time-ordered data to predict future values.

Trend (Time-Series)

Long-term upward or downward movement in data due to factors like population or income changes.

Seasonality

Short-term, regular demand fluctuations tied to calendar or time-of-day events (e.g., weekends, holidays).

Cycle

Wavelike variations in data lasting more than one year, often linked to economic or political conditions.

Irregular Variation

Unusual, non-recurring events (strikes, extreme weather) causing data deviations that are not typical.

Random Variation

Residual, unexplained variability remaining after other patterns are accounted for.

Naïve Forecast

Forecast that sets next period’s value equal to the most recent actual value; simplest time-series method.

Moving Average

Forecast based on the average of a specified number of the most recent actual values.

Weighted Moving Average

Moving average giving more importance (weight) to recent observations than older ones.

Exponential Smoothing

Weighted averaging method that updates forecast by adding a fraction (alpha) of the latest forecast error.

Linear Trend Equation

Straight-line model (y = a + bx) fitted to historical data to project future values.

Trend-Adjusted Exponential Smoothing

Exponential smoothing method that includes a separate trend factor (beta) to track upward or downward movement.

Additive Seasonality Model

Approach that adds or subtracts a constant seasonal amount to a trend or average.

Multiplicative Seasonality Model

Approach that multiplies trend or average by a seasonal percentage (seasonal relative).

Seasonal Relative

Seasonal index expressed as a percentage used to adjust forecasts in multiplicative seasonality models.

Deseasonalizing Data

Removing seasonal effects by dividing each observation by its seasonal relative to reveal underlying trend.

Associative (Causal) Forecasting

Building an equation relating the variable of interest to one or more predictor variables.

Predictor Variable

Independent variable used in an associative model to estimate the dependent (forecast) variable.

Simple Linear Regression

Technique fitting a straight line between one independent and one dependent variable using least squares.

Least Squares Criterion

Process of minimizing the sum of squared vertical deviations between data points and the regression line.

Correlation Coefficient (r)

Statistic measuring strength and direction of linear relationship; ranges from −1.00 to +1.00.

Coefficient of Determination (r²)

Proportion of variability in the dependent variable explained by the independent variable; ranges 0–1.

Linear Regression Assumptions

1) Random variation around line, 2) Normally distributed errors, 3) Predictions limited to observed data range.

Control Chart (for Forecast Errors)

Graphical tool with upper and lower limits used to detect non-random forecast errors and model bias.

Tracking Signal

Ratio of cumulative forecast error to MAD; used to identify persistent bias in forecasts.

Factors in Choosing a Forecasting Technique

Cost, accuracy, data availability, software availability, time to prepare, and forecast horizon.

Operations Strategy (Forecasting)

Using better forecasts to exploit opportunities, reduce risks, shorten horizons, share data, and improve supply-chain decisions.