A fuzzy system model for financial assessment of listed companies

This paper presents an approach to the problem of listed companies´ assessment. The valuation depends on the listed companies finance. The most direct financial picture of a company is given by its financial statements, which are published, according to the relative legislation, periodically. The interpretation of financial statements involves high degree of subjectivity, therefore a fuzzy logic approach, which is closer to human logic, is more suitable. Based on the anfis matlab platform, we build the fuzzy assessment model by combining three Mamdani Fuzzy Inference System (FIS) and one Takagi-sugeno (TS) FIS, which operates based on financial indices values. The purpose of this composition lays on the fact that the Mamdani exports fuzzy inferences more closely to the human logic. On the other hand, the TS model, which provides the final output, can be trained for more accurate results. The model can be used from stakeholders and possible investors in order to assess the risk level through categorization of financial dangerous companies and healthy ones.