Explanation generation in business performance models - With a case study in competition benchmarking

In this paper, we describe an extension of the methodology for explanation generation in financial knowledge-based systems. This offers the possibility to automatically generate diagnostics to support business decision tasks. The central goal is the identification of specific knowledge structures and reasoning methods required to construct computerized explanations from financial data and models. A multi-step look-ahead algorithm is proposed that deals with so-called cancelling-out effects. The extended methodology was tested on a case-study conducted for Statistics Netherlands involving the comparison of financial figures of firms in the Dutch retail branch. The analysis is performed with a diagnostic software application which implements our theory of explanation. Comparison of results of the method described in (Daniels and Feelders, 2001) with the results of the extended method clearly improves the analyses when cancelling-out effects are present in the data.