Site MapHelpFeedbackGlossary
Glossary
(See related pages)


alpha (α)  0 < α < 0.05 The smoothing constant in exponential smoothing.
associative forecasting  forecasting such as regression that estimates the effects of a causal variable or variables upon a dependent variable.
bias  The tendency of forecast errors to be successively positive or negative.
centered moving average  a moving average that is positioned at the middle of the group of numbers it represents.
control chart  a technique for monitoring the accuracy of forecasts. Acceptable forecast errors fall within control limits which are multiples of the MSE. Control charts are used heavily in quality control and are discussed at greater length in Chapter 10.
control limits  the outer limits (positive and negative) of acceptable values of the tracking signal. Values outside of the control limits indicate problems with the forecasts.
correlation  the degree to which the values of two variables have a tendency to change in the same (or opposite) directions.
cycles  recurrent fluctuations up and down in a time series occurring over a period of years.
decomposition  isolating seasonal cyclical, trend and random effects in a time-series data.
Delphi technique  forecasting that forms a consensus of qualified experts, in a series of stages; the results of the questionnaire at one stage are used to redesign the questionnaire at the next stage, and, if necessary, to redefine the problem for which the predictions are being made.
dependent variable  the variable whose value is to be predicted or estimated; symbolized by y.
double exponential smoothing (trend-adjusted exponential smoothing)  removing the lag inherent in single exponential smoothing, by a model that combines single exponential smoothing of the raw data with secondary exponential smoothing of the predictions.
error  the difference between the actual value and the value that was forecasted for the same time.
exponential smoothing  the forecasted value consists of the previous period's forecast plus a percentage of the forecast error.
Forecast  a statement about the future.
irregular variation  variation caused by unusual circumstances.
judgmental forecast  a forecast based entirely upon the experience and personal opinions of a qualified expert.
lag  a specified difference between actual and forecasted values, expressed in time units (e.g., years), rather than in customary physical or monetary units.
least squares line  the line which minimizes the sum of the squared deviations of the data points of the dependent variable from the line.
linear regression  see regression.
linear trend line  the least squares line which has time as the causal variable.
mean absolute deviation (MAD)  the average of the absolute differences between the actual values and the predictions.
mean square error (MSE)  the variance computed by averaging the squared differences between the predicted and the actual values.
moving average  a sequence of averages, derived from a set of time-series data, for a period of time, such as a season, week, month or day; each average covers n such periods and is computed by deleting the first period from the previous MA, and adding in the most recent period.
naive forecast  a prediction that the value in the next period will be the same as what it is in the current period, or else that the difference between the current period and the next one will be the same as the difference between the current period and the most recent one.
predictor variable  variable that can be used to predict values of the variable of interest; symbolized by x.
random variation  residual variation after all other factors are accounted for.
regression  a technique for generating forecasts by computing a linear (straight-line) equation that shows the influence of a causal variable upon the dependent variable whose values are being forecasted.
seasonality  fluctuations up or down in a time series which recur regularly, often annually.
seasonal relative  an index that expresses the degree to which the values of a time series in a particular season tend to vary from the "average" if there were no seasonal variations.
time series  a time-ordered sequence of observations, made at regular intervals.
tracking signal (TS)  a measure of the difference between the forecasted and the actual values; a good forecast will always have a TS within a range centering on zero, for example ±4, and there would be about the same number of values above zero as below zero.
trend  a long-term tendency of a time series to increase or decrease.
weighted average  more recent values in a time series are given more weight in computing a forecast.







Stevenson OM7Online Learning Center with Powerweb

Home > Chapter 3 > Glossary