Operations Management: An Integrated Approach, 5th Edition

SELECTING THE RIGHT FORECASTING MODEL

A number of factors influence the selection of a forecasting model. They include the following:

1. Amount and type of available data. Quantitative forecasting models require certain types of data. If there are not enough data in quantifiable form, it may be necessary to use a qualitative forecasting model. Also, different quantitative models require different amounts of data. Exponential smoothing requires a small amount of historical data, whereas linear regression requires considerably more. The amount and type of data available play a large role in the type of model that can be considered.
2. Degree of accuracy required. The type of model selected is related to the degree of accuracy required. Some situations require only rough forecast estimates, whereas others require precise accuracy. Often, the greater the degree of accuracy required, the higher is the cost of the forecasting process. This is because increasing accuracy means increasing the costs of collecting and processing data, as well as the cost of the computer software required. A simpler and less costly forecasting model may be better overall than one that is very sophisticated but expensive.
3. Length of forecast horizon. Some forecasting models are better suited to short forecast horizons, whereas others are better for long horizons. It is very important to select the correct model for the forecast horizon being used. For example, a manufacturer that wishes to forecast sales of a product ...