Forecasting and Prediction

Abstract

This chapter discusses the avenues by which decision-makers attempt to build predictive models that afford accurate prediction of their chosen criteria. Probably the best example of using models on which to base predictions is weather forecasting. The data, the context, and the budget determine the viability of quantitative forecasting. Artificial neural networks rely on the fitting of testing data to make predictions on new data. The most basic and popular associative model is linear regression, which tries to fit a line to the data. Quantitative methods entail judgment, to select one method or aspects of a method, to make assumptions about the data-generating process. Making assumptions explicit facilitates the identification of possible errors made by experts within a given forecast. Moreover, the transparency of assumptions is also helpful for predictions that rely on the information provided by consumers about their own attitudes, preferences, or behaviors.