A Hybrid intelligent system for diagnosing and solving financial problems

Our main purpose in this dissertation is to develop a system to diagnose and indicate solutions to financial health problems of small and medium firms (SMF). Although monitoring and adjusting financial problems play a central role in the firm’s performance, usually SMF face difficulties in these tasks for lacking human resources and for having incapacity to afford a consultant. The closest types of systems available in the literature are the bankruptcy prediction, the credit analysis and the auditing models. Bankruptcy models do not work for the purpose of financial health evaluation because, rather than looking for causes and corrections to deviations, they intend to foresee the death or life of the firm. Credit analysis models are developed to creditors interested only in the safety of their investments. Auditing systems are limited to ratio analysis with general comments about the financial condition of a firm. After all, the critical aspect of offering practical solutions to the firm remains open. There is need for a financial advisor that helps the manager to make financial decisions that lead to a long-run profitability and success of the firm. A study of financial statement analysis shown that there are two different reasoning processes participating of the solution: inductive and deductive. We implemented a hybrid intelligent (neuro-fuzzy-symbolic) system to combine the inductive and deductive reasoning processes. The inductive reasoning was modeled by the connectionist module while the deduction was implemented through a fuzzy expert system.