No abstract

No abstract

The purpose of this study was to determine which forecasting method would most accurately predict participation in congregate lunch programs serving older adults, and whether forecasts would improve if the historical information used were obtained from the same day of the menu cycle as opposed to the same day of the week. Time‐series forecasting techniques including naïve, moving average, and simple exponential smoothing were applied to data collected over a 6‐month period from eight meal sites. In all the cases studied where the same day of the week was used as the basis for the historical information, exponential smoothing with the best alpha outperformed both the Naïve and the Moving Average (N = 3) methods. Analyses using paired t‐tests revealed that there were no significant improvements in forecasting when the same day of the menu cycle, as opposed to the same day of the week, was used as the source for historical data.

The purpose of this study was to determine which forecasting model would most accurately predict meal participation at university residential dining facilities. Forecasting techniques including naive, moving average (3 versions), and simple exponential smoothing were applied to data collected from two dining halls over 15 weeks. An analysis of the forecasting models using Mean Absolute Deviations (MAD), Mean Squared Errors (MSE), and Mean Absolute Percentage Errors (MAPE) indicated that simple mathematical forecasting techniques provided better predictions (MAPE < 6.4%) than the naive method in all cases studied. Moving average methods outperformed other methods 83% of the time. Implications are discussed.

No abstract