Forecasting techniques

Computer-assisted forecasting techniques have been presented. Time-series extrapolation techniques are based upon the historical pattern of the item to be forecasted. Of particular importance are causal (functional) models which take influencing factors into consideration. Regardless whether the causal model comprises a single relation or multiple relations, one must be aware of the various tests and evaluating methods in the stages of model building in order to obtain better forecast models. Building and analyzing a forecasting model normally requires data manipulation and calculation. Unless these tasks are performed in a systematic manner, a heavy burden is placed on the model-builder. Productivity in forecasting activities is increased by using a systematic computer programming system oriented for forecasting. Introduced as an example forecasting program was FAMS. Further important steps forward are flexible linkage of techniques, linkage of models, and linkage of planning applications. By realizing these, forecasting will lay a foundation of management science applications and provide a framework for timely and sound decision making.