Incorporated intangible assets with a multiple-agent decision tree for financial crisis prediction

With the current financial scandals and European debt crisis, corporate financial crisis prediction has become an essential task in the fields of financial and risk management. Numerous pre-warning mechanisms based on statistical or artificial intelligence theories have been introduced in the literature, yet no current pre-warning model presents the best performance under all measurements. With significant great improvements in information technologies and computational techniques, the multi-agent mechanism or ensemble learning has been proposed as an efficient way to achieve superior performance. The fundamental concept of multi-agent learning aims at complementing errors made by a singular method, and thus this study proposes a pre-warning model based on multi-agent learning and further considers the impacts from intangible assets, which are at the core of value-creating procedures in a knowledge economy, on constructing such a model. This model can help firms with a large amount of intangible assets to have a higher possibility of gaining considerable wealth in the future and a lower possibility for encountering financial troubles. The introduced pre-warning mechanism is a promising alternative for predicting financial crises, is supported by real cases, and assists managers to modify their capital structure and debt leverage.