6. Conclusion
Uncertainty appears in different forms and. affects decision making. Nowadays, there are mathematical models to handle the uncertainty. But if we work with a knowledge base that changes with time, and with non-contradictory information that becomes doubtful or contradictory, or with any combination of these three situations then we need to use mediative fuzzy logic which is able to process inconsistent information.
So is mentioned in first chapter, applications based on mediative fuzzy logic (see papers [17–22]) have shown its superiority to other fuzzy logics (traditional or intuitive). For this reason, in this paper, we extends and improves the system from [30] based on fuzzy logic by working with intuitionistic fuzzy sets to represent the input and output variables and with mediative fuzzy logic for reasoning. Superiority of our system is given by the possibility to handle contradictory and doubtful information. As is mentioned in [22] the fact of having the possibility of complementing the knowledge with non-agreement functions give us the possibility of implementing a more realistic fuzzy inference system.
In future papers we intend to improve this system by
• tuning the membership functions and rules used in inference system; for instance, if-then rules can be obtained from training patterns
• using a procedure to generate inference rules so that each has a degree of certainty
• improving the reasoning system by using the degree of certainty for determining the inferred conclusion
• test with other equations to compute the mediate output; for instance (8) and (9).
