Analysis of the inconsistency problem in the model for predicting "diseases" of firms

The development of economic-financial diagnostic models, using fuzzy logic tools, has resulted in important applications. It is possible to identify a matrix of economic-financial knowledge (Matrix R), which simulates the analyst's behaviour evaluating the health of a firm. The proposed models, from a set of symptoms and causes, use a methodology of equations resolution in fuzzy relations. In particular, we have applied their inverse problem. However, it is possible to identify some inconsistencies in the application of the resolution method, which could invalidate several elements of the matrix R. In this paper, we discuss these inconsistencies and propose a method to eliminate them. So we can find a matrix R where all its elements represent the true relationship between symptoms and causes.