Deriving domain knowledge from unstructured information: A forecasting framework based on judgmental adjustment approach

Abstract

The integration of quantitative and judgmental forecasting methods have been increasingly applied to give better performance to forecast. Judgmental adjustment is one instance of integrating both methods and it has been gaining recognition among forecasting practitioners because of its quick and convenient way to perform forecast. However, many criticize this approach because of its disadvantages, i.e. bias and inconsistency which are associated to the human. We are proposing a forecasting framework that aids the process of judgmental adjustment by providing supportive information to reduce the effect of bias and inconsistency. The proposed framework comprises five different modules, i.e. time series graphical display, quantitative forecast, news-based supportive information, user comment and similaritybased pattern search.