- مبلغ: ۸۶,۰۰۰ تومان
- مبلغ: ۹۱,۰۰۰ تومان
Bankruptcy prediction models that rely on ensemble techniques have been studied in depth over the last 20 years. Within most studies that have been performed on this topic, it appears that any ensemble-based model often achieves better results than those estimated with a single model designed using the base classifier of the ensemble, but it is not uncommon that the results of the former model do not outperform those of a single model when estimated with any other classifier. Indeed, an ensemble of decision trees is almost always more accurate than a single tree but not necessarily more than a neural network or a support vector machine. We know that the accuracy of an ensemble used to forecast firm bankruptcy is closely related to its ability to capture the variety of bankruptcy situations. But the fact that it may not be more efficient than a single model suggests that current techniques used to handle such a variety are not completely satisfactory. This is why we have looked for a method that makes it possible to better embody this diversity than current ones do. The technique proposed in this article relies on the quantification, using Kohonen maps, of temporal patterns that characterized the financial health of a set of companies, and on the use of an ensemble of incremental size maps to make forecasts. The results show that such models lead to better predictions than those that can be achieved with traditional methods.
In this study, we propose a new bankruptcy prediction model that relies on the estimation of failure patterns that are quantified with ensembles of Kohonen maps. The results show that this type of model is, on the whole, more efficient than single or ensemblebased models and this efficiency seems to be rather robust since it was assessed using different samples collected over different time periods. Moreover, failure pattern-based models embody a predictive function that appears to be more complex than that embodied by any type of model since the bias component of their error is markedly lower than that of all other models. One may then think that their performance is likely due to this complexity.
Our findings have two main implications. First, they reinforce the idea that emerged in the literature a few years ago and which suggests that model accuracy does not solely rely on data mining techniques but also on the way one will use some knowledge about the bankruptcy phenomenon during a modeling process; incorporating domain knowledge into classification models may indeed improve model performance . Here, domain knowledge relies on the concept of “failure patterns”, which has been highlighted in the literature that deals with financial or organizational issues, but also on the limits of the techniques that have been used so far to assess these patterns. Thus, a modeling framework can really benefit from conceptual developments that make it possible to enrich the knowledge that one can have about a firm failure. Our findings invite the scientific community that is interested in this issue to renew the conceptual basis used to design models. Second, they remind financial institutions of the true added value of ensemble-based models by once again showing them the limits of their own models and by providing them with a reliable modeling framework that may fit their needs.